Virulence determinant, PTP7, controls vesicle budding from the Maurer’s clefts, adhesin protein trafficking and host cell remodeling in Plasmodium falciparum

Presentation of the variant antigen, Plasmodium falciparum erythrocyte membrane protein 1 (EMP1), at knob-like protrusions on the surface of infected red blood cells, underpins P. falciparum malaria pathogenicity. Here we describe a protein PF3D7_0301700 (PTP7), that functions at the nexus between the intermediate trafficking organelle, the Maurer’s cleft, and the infected red blood cell surface. Genetic disruption of PTP7 leads to accumulation of vesicles at the Maurer’s clefts, grossly aberrant knob morphology, and failure to deliver EMP1 to the red blood cell surface. We show that an expanded low complexity sequence in the C-terminal region of PTP7, found only in the Laverania clade of Plasmodium, is critical for efficient virulence protein trafficking. Author Summary We describe a malaria parasite protein involved in virulence factor trafficking (PTP7) that moves between different compartments in the host red blood cell cytoplasm in a stage-dependent manner. Upon disruption of the PTP7 locus, the Maurer’s cleft trafficking compartments become decorated with vesicles; the knobby protrusions on the host red blood cell surface are depleted and distorted; and trafficking of the virulence protein, EMP1, to the host red blood cell surface is ablated. We provide evidence that a region of PTP7 with low sequence complexity plays an important role in driving fission of vesicles from the Maurer’s clefts.


Introduction 32
Plasmodium falciparum causes more than 200 million malaria infections every year, killing more 33 than 400,000 people [1]. Central to the ability of P. falciparum to maintain an infection and cause 34 disease, is the invasion and remodelling of host red blood cells (RBCs). Maturation of the parasite 35 inside the RBC is accompanied by striking changes in the surface topology of the infected RBC and 36 a marked loss of cellular deformability [2,3]. One key modification is the assembly of ~90 nm 37 diameter structures, called knobs, at the infected RBC surface [4]. The knob structure acts as an 38 elevated platform at the RBC surface for presentation of the major virulence protein P. falciparum 39 erythrocyte membrane protein-1 (EMP1), which mediates binding of infected RBCs to endothelial 40 ligands [5][6][7][8]. 41 The trafficking of EMP1 and other virulence determinants beyond the confines of the parasite is 42 mediated by the Plasmodium translocon of exported proteins (PTEX) that is present in the 43 parasitophorous vacuole membrane (PVM), with help from a second complex termed the exported 44 protein interacting complex (EPIC) [9][10][11][12]. Once exported across the PVM, EMP1 is thought to be 45 amounts of EMP1 being trafficked to the Maurer's clefts had changed we performed a quantitative 165 analysis of the images. We looked at the co-occurrence of EMP1 and the Maurer's clefts marker, 166 REX1. A significant increase in the number of EMP1-containing clefts (Fig 3E, S8 Fig) and a total  167 increase in the amount of EMP1 at the clefts was observed in the knockout compared to the CS2 168 parasites (Fig 3G, S8 Fig). Additionally, this analysis revealed that the ∆PTP7-infected RBCs exhibited 169 fewer and larger clefts than CS2 controls (Fig 3H, I). 170

Vesicles accumulate at the Maurer's clefts in PTP7-disrupted parasites 171
Ultrastructural analysis of the Maurer's clefts was performed in infected RBCs that had been 172 permeabilized with Equinatoxin II to allow introduction of antibodies [16]. In wildtype parasites, the 173 Maurer's clefts are observed as single slender cisternae with an electron-lucent lumen and an 174 electron-dense coat. Upon immuno-labelling with anti-REX1, gold particles are observed at the cleft 175 periphery (Fig 4A top,  magenta arrows). These data suggest that these budding structures may be EMP1-trafficking 188 vesicles that fail to separate from the clefts in the absence of PTP7. 189

The C-terminal asparagine repeats are needed for PTP7 function 190
The majority of the PTP7 sequence is highly conserved across Plasmodium species, however within 191 P. falciparum isolates and one clade of P. gorilla there has been the insertion of 34 asparagines near  Scanning electron microscopy was performed on the truncations to determine if the knob 203 morphology was affected. Truncation of the C-terminal 17 amino acids (PTP7 ∆300-317 ) a minor effect 204 on knob density and knob diameter (Fig 5C-E). By contrast, deletion of the entire C-terminal region 205 (PTP7 ∆265-317 ) was associated with fewer, larger knobs ( Fig 5C-E), recapitulating the phenotype 206 observed in the PTP7 knock-out line. Interestingly, internal deletion of the C-terminal asparagine 207 repeats, while keeping the C-terminal 'KKSKKN' motif (PTP7 ∆278-310 ), causes an intermediate 208 phenotype. The PTP7 ∆278-310 parasites exhibit sparser knobs with larger diameters (Fig 5C-E). 209 We next investigated if delivery of EMP1 to the RBC surface was affected in the truncations. Flow 210 cytometry analysis revealed complete ablation of EMP1 at the surface of the PTP7 ∆265-317 infected 211 RBCs ( Fig 6A). The PTP7 ∆278-310 infected RBCs had a 63% reduction in surface EMP1 while the and 212 PTP7 ∆300-317 infected RBCs had a 32% reduction (Fig 6A). A trypsin cleavage assay was performed to 213 confirm these results. Complete loss of EMP1 from the surface is observed in the PTP7 ∆265-317 214 infected RBCs, while faint trypsinized EMP1 bands are observed in both PTP7 ∆278-310 and PTP7 ∆300-317 215 parasites, consistent with the reduction of surface-exposed EMP1 detected in the FACS experiments 216 ( Fig 6B). In a similar trend, the number of GFP-positive puncta is significantly reduced in PTP7 ∆265-317 217 and PTP7 ∆278-310 parasites (Fig 1C, D, S10 Fig). These puncta counts suggest that the length of the C-218 terminal region of PTP7 affects its distribution. 219 To determine the effect of truncation of regions in the C-terminal domain of PTP7 on vesicular 220 trafficking, thin section TEM was performed. Truncation of the entire C-terminal region leads to 221 marked vesiculation of the Maurer's clefts as revealed in ultrastructural analysis ( Fig 6C). The 222 Maurer's clefts appear less affected when at least some of the PTP7 C-terminal domain is retained 223 (PTP7 ∆278-310 and PTP7 ∆300-317 ; Fig 6C). The vesicle diameters are not significantly affected by PTP7 224 mutations ( Fig 6D); however, significantly more vesicles are observed within 100 nm of the PTP7 ∆265-225 317 clefts, while the PTP7 ∆278-310 and PTP7 ∆300-317 parasites exhibited a non-significant increase, 226 relative to the full-length control ( Fig 6E). The accumulation of vesicles at the clefts is similar in the 227 PTP7 ∆265-317 and ∆PTP7 lines. 228

Discussion 229
The canonical protein trafficking system that is used by most eukaryotic cells to transport proteins 230 to the plasma membrane is not present in mature human RBCs [40]. Thus, intraerythrocytic P. 231 falciparum faces the challenge of exporting integral membrane proteins, such as the virulence 232 antigen EMP1, to the RBC membrane, and inserting proteins into the lipid bilayer, in the absence of 233 To assess the function PTP7 plays in virulence protein trafficking, we generated a PTP7 knockout cell 252 line. Analysis of this cell line showed that PTP7 is not essential for growth of parasites in culture, in 253 agreement with a previous report [47]. However, in the absence of PTP7, EMP1 becomes stuck at 254 the Maurer's clefts and is not presented at the surface of the infected RBC. Interestingly, upon 255 deletion of PTP7, the Maurer's clefts became decorated with numerous vesicle-like structures that 256 appear to be in the process of budding from the cleft surface. Quantitative imaging revealed an 257 accumulation of EMP1 at the Maurer's clefts in the ∆PTP7 infected RBCs. Immuno-EM, using an 258 antibody to EMP1, reveals labeling of both vesicles and the Maurer's clefts. Taken together, this 259 suggests that the vesicles may be in the process of budding from the clefts, and that PTP7 is required 260 for correct vesicle fission and forward trafficking of EMP1 to the RBC membrane. 261 In addition to accumulation of vesicles at the clefts, the PTP7 knockout parasites also exhibit altered 262 knob morphology, with fewer, larger knobs being assembled at the RBC surface. The main structural 263 component of the knob complex, KAHRP, has been shown to traffic to the RBC membrane directly; 264 that is, not via the Maurer's cleft. Our data for the PTP7 knockout suggests convergence of the 265 KAHRP and EMP1 trafficking pathways during the final step, where EMP1 is loaded into the knobs 266 at the RBC membrane. One possibility is that the vesicles also transport a component needed for 267 correct assembly of knobs. 268 It is interesting to consider how PTP7, which has no ATP or GTP binding domain, could provide the 269 driving force for vesicle fission. In this context, the unusual asparagine repeat sequence in the C- To test this hypothesis and to examine the role of the PTP7 asparagine repeats in the molecular 277 architecture of the infected RBC and in the EMP1 trafficking process, we generated transfectants in 278 which PTP7 was truncated. Complete removal of the C-terminal domain recapitulated the 279 phenotype seen in ∆PTP7; that is, numerous vesicles accumulate at the Maurer's clefts, while the 280 knobs showed aberrant morphology and EMP1 trafficking is ablated. The Maurer's cleft phenotype 281 was less dramatic if just the C-terminal basic motif was removed or if the asparagine repeats were 282 removed but the C-terminal polybasic motif was maintained. However, in all truncation mutants, 283 the trafficking of EMP1 to the surface was decreased suggesting that the length of the C-terminal 284 region also contributes to efficient EMP1 trafficking. 285 Conservation of PTP7 in the Laverania subgenus may be linked to the expansion of virulence protein 286 trafficking and/or host cell remodeling machinery in these species. The N-terminus of PTP7 is well 287 conserved in gorilla, gaboni and reichenowi orthologues, but the asparagine repeats are found only 288 in P. falciparum and a clade of P. gorilla [50]. This expansion of the PTP7 sequence may be associated 289 with increased virulence, which is supported by our truncation data showing that this region is 290 required for EMP1 trafficking. 291 Taken together, our data provide strong evidence that PTP7 and its interacting proteins play a 292 critical role in the budding of EMP1-containing vesicle-like structures from the Maurer's clefts and 293 their transfer to the host RBC membrane. Our data provide intriguing evidence for a key role of a 294 low complexity region in providing the driving force for vesicle fission. The promiscuous localization 295 and protein-protein interactions of PTP7 suggest that the routes of exported protein trafficking may 296 be more integrated and interdependent than previously thought. The suggestion that the vesicles 297 that bud from the Maurer's clefts contain an unidentified component that is needed for correct 298 assembly of the knob complex needs further exploration. An increased understanding of the 299 processes for trafficking virulence proteins may lead to new therapies to tackle malaria 300 pathogenesis. 301

Materials and Methods 302
Ethics statement. Red blood cells and serum were acquired from the Australian Red Cross Lifeblood 303 blood service. All blood products were anonymous and individual donors could not be identified. 304 Vector construction and generation of transgenic parasites. For endogenous 3' tagging, the 3' 710 319 bp of the ptp7 was amplified using SLI-sand-PTP7_fw and SLI-sand-PTP7_rv primers (S1 Table) and 320 directionally cloned into the NotI/AvrII restriction sites in the pSLI-sandwich plasmid (2xFKBP-GFP-321 2xFKBP) [27]. Ring-stage parasites were transfected with 50-100 µg of plasmid as previously 322 described [55]. Briefly, precipitated plasmid was resuspended in sterile TE buffer and Cytomix. 5-323 10% ring-stage infected RBCs were resuspended in the DNA mix and transferred to a 2 mm 324 electroporation cuvette (BTX). Cells were electroporated at 310 V, 950 µF, and ∞ resistance (Gene 325 Pulser Xcell™ Electroporation System, Bio-Rad), then washed in warm RPMI media and transferred 326 to culturing dishes. After neomycin selection, correct integration of the plasmid was verified by PCR 327 using the primers as described in S1 Fig and S1 Table.  328 To generate the plasmid for CRISPR/Cas9 mediated disruption of ptp7, 5' and 3' homology regions 329 (HR; ~500 bp) were PCR-amplified from genomic DNA using the primers PTP7-HR1_fw, PTP7-HR1_rv 330 for the HR1 and PTP7-HR2_fw, PTP7-HR2_rv for HR2 (S1 Table). The HR1 fragment was cloned into 331 the AvrII/NcoI sites and the HR2 into the SpeI/SacII sites of the pUF-TK plasmid [56]. The pUF-TK 332 vector was linearized by digestion with AvrII and used as the repair template. CRISPR/Cas9-333 mediated double-stranded breaks were guided by a single guide RNA and Cas9. The Cas9 target 334 'ggttccaacacagtcacacg' was selected using CHOPCHOP and PTP7-sgRNA_top and PTP7-335 sgRNA_bottom oligonucleotides were annealed and cloned into the BsrgI site of pAIO using infusion 336 cloning [56]. Both linearized repair template and guide plasmids were transfected simultaneously 337 into CS2 parasites. Gene disruption was confirmed by PCR using primers yDHODH_ScrF and 338 yDHODH_ScrR, and primers in the native 5' and 3' UTRs of the ptp7 locus, PTP7-KO_ScrF and PTP7-339 KO_ScrR (Fig 2, S1 Table). 340 To generate the truncation parasite lines the DNA sequences encoding for amino acids 83-317, 83-341 264, 83-277, and 83-299 of PTP7 were PCR-amplified using the primers indicated (S1 Table)