Co-immunoprecipitation with MYR1 identifies three additional proteins within the Toxoplasma parasitophorous vacuole required for translocation of dense granule effectors into host cells

Toxoplasma gondii is a ubiquitous, intracellular protozoan that extensively modifies infected host cells through secreted effector proteins. Many such effectors must be translocated across the parasitophorous vacuole (PV) in which the parasites replicate, ultimately ending up in the host cytosol or nucleus. This translocation has previously been shown to be dependent on five parasite proteins: MYR1, MYR2, MYR3, ROP17, and ASP5. We report here the identification of several MYR1-interacting and novel PV-localized proteins via affinity purification of MYR1, including TGGT1_211460 (dubbed MYR4), TGGT1_204340 (dubbed GRA54) and TGGT1_270320 (PPM3C). Further, we show that three of the MYR1-interacting proteins, GRA44, GRA45, and MYR4, are essential for the translocation of the Toxoplasma effector protein GRA16, and for the upregulation of human c-Myc and cyclin E1 in infected cells. GRA44 and GRA45 contain ASP5-processing motifs, but like MYR1, processing at these sites appears to be nonessential for their role in protein translocation. These results expand our understanding of the mechanism of effector translocation in Toxoplasma and indicate that the process is highly complex and dependent on at least eight discrete proteins. Importance Toxoplasma is an extremely successful intracellular parasite and important human pathogen. Upon infection of a new cell, Toxoplasma establishes a replicative vacuole and translocates parasite effectors across this vacuole to function from the host cytosol and nucleus. These effectors play a key role in parasite virulence. The work reported here newly identifies three parasite proteins that are necessary for protein translocation into the host cell. These results significantly increase our knowledge of the molecular players involved in protein translocation in Toxoplasma-infected cells, and provide additional potential drug targets.

Toxoplasma gondii is a ubiquitous, intracellular protozoan that extensively 25 modifies infected host cells through secreted effector proteins. Many such effectors 26 must be translocated across the parasitophorous vacuole (PV) in which the parasites 27 replicate, ultimately ending up in the host cytosol or nucleus. This translocation has 28 previously been shown to be dependent on five parasite proteins: MYR1, MYR2, MYR3, 29 ROP17, and ASP5. We report here the identification of several MYR1-interacting and 30 novel PV-localized proteins via affinity purification of MYR1, including TGGT1_211460 31 (dubbed MYR4), TGGT1_204340 (dubbed GRA54) and TGGT1_270320 (PPM3C). 32 Further, we show that three of the MYR1-interacting proteins, GRA44, GRA45, and 33 MYR4, are essential for the translocation of the Toxoplasma effector protein GRA16,34 and for the upregulation of human c-Myc and cyclin E1 in infected cells. GRA44 and 35 GRA45 contain ASP5-processing motifs, but like MYR1, processing at these sites 36 appears to be nonessential for their role in protein translocation. These results expand 37 our understanding of the mechanism of effector translocation in Toxoplasma and 38 indicate that the process is highly complex and dependent on at least eight discrete 39 proteins. 40

Introduction 51
Toxoplasma gondii is an obligate intracellular parasite that can cause severe 52 illness in immunocompromised individuals and the developing fetus. It is estimated to 53 infect up to a third of the world's population, and has an unparalleled host range, 54 infecting virtually any nucleated cell in almost any warm-blooded animal (1). In order to 55 survive within a host cell, Toxoplasma tachyzoites, the rapidly-dividing, asexual stage of 56 the parasite, establish a replicative niche, the parasitophorous vacuole (PV), whose 57 membrane (PVM) acts as the interface between parasite and host. While the PV 58 protects intracellular Toxoplasma from clearance by the innate immune system, it also 59 acts as a barrier that Toxoplasma must overcome in order to hijack host resources. 60 Toxoplasma extensively modifies the host cells it infects via secreted effectors, 61 acidic pI of 4.91 (ToxoDB v45) which is known to reduce protein mobility on SDS-PAGE 172 (33), and/or to post-translational modifications (all three proteins are reported to be 173 phosphorylated (31; ToxoDB v45)). This same slower-than-expected mobility for the 174 major band was seen for an independently generated, cloned line expressing HA-175 tagged 211460 (Fig. S1A) and so we conclude that this is the correct mobility for this 176 protein. Interestingly, both the 211460-3xHA tagged population and single clone also 177 showed a smaller but considerably weaker band at around the expected size (~100 178 kDa). Whether this smaller MYR4 product is biologically relevant, or is simply a product 179 of protein degradation, is unclear. PV-localization for 211460 is further confirmed in the independently generated clonal 188 line (Fig. S1B). Thus, we conclude that 211460, 204340, and PPM3C are at least 189 transiently localized to the Toxoplasma PV during infection. Furthermore, we also 190 assessed the localization of these proteins within the parasites themselves. The results 191 ( Fig. S2) show that while PPM3C appears to be present throughout the parasite, 192 with the dense granule protein GRA7, suggesting that these two proteins are also GRA 194 proteins. We therefore designate 204340 as GRA54 for its GRA-like localization, and 195 211460 as MYR4, for reasons described below. 196 To assess their potential involvement in GRA effector translocation, we 197 attempted to generate knockouts of our candidate genes in a strain of Toxoplasma that 198 constitutively expresses an HA-tagged version of the MYR1-dependent secreted 199 effector protein GRA16, RH∆gra16::GRA16-HA ("parental"). To do this, we co- integration of the vector by PCR with gene-specific primers. Using this strategy, we 205 were able to disrupt the genomic loci of MYR3, GRA44, GRA45, CST1, GRA54, MYR4, 206 PPM3C, and GRA7 (Fig. S3). Despite several attempts, however, we were unable to 207 generate a GRA52 mutant. This gene may be essential as it has a very negative 208 CRISPR fitness score of -3.96 (35). Given that the MAF1 locus is expanded in 209 Toxoplasma, with 4 copies in RH parasites (36), we chose not to attempt a 210 CRISPR/Cas9 approach to knockout MAF1a, and instead utilized a previously 211 generated strain in which the entire MAF1 cluster (including MAF1a and MAF1b) is 212 To determine if the absence of any of the candidate genes results in a defect in 214 effector translocation across the PVM, we used IFA to assess both GRA16-HA export to 12 infection (37)) in the disrupted lines. Quantified results for all nine genes tested show 217 that disruption of GRA44, GRA45, MYR4 and the previously described MYR3, all 218 resulted in a complete or near-complete block in GRA16 export to the host nucleus 219 (Figs. 3B, S4) and a failure to upregulate host c-Myc (Figs. 3C, S4); on the other hand, 220 disruption of GRA7, CST1, GRA54, or PPM3C resulted in no detectable effect on either 221 of these two phenotypes. Additionally, we found that the previously generated ∆maf1 222 strain also had normal GRA16 export to the host nucleus (Fig. 3D). These results 223 indicate that of the nine genes tested here, only MYR3, GRA44, GRA45 and MYR4 are 224 necessary for the translocation of GRA effectors across the PVM. 225 To test the generality of their role in effector translocation, we next assessed the 226 impact of these gene disruptions on the upregulation of host cyclin E1 which has been 227 shown to be dependent on export of the MYR1-dependent effector HCE1/TEEGR (6). 228 The results showed that, as for GRA16, disruption of MYR3, GRA44, GRA45, and 229 MYR4 also resulted in a block in cyclin E1 upregulation in infected host cells, while no 230 obvious defect was observed in the parasite lines disrupted in GRA7, CST1, GRA54 231 and PPM3C (Fig. 3E). A repetition of the cyclin E1 western blot with higher parasite 232 input reveals that the absence of cyclin E1 upregulation observed in ∆gra44 parasites in 233 Fig. 3E is not due to low parasite input in that particular experiment (Fig. S5). These 234 results argue that GRA44, GRA45 and MYR4 are all required for translocation across 235 the PVM of at least two independent GRA effectors. 236 Our previous work has shown that deletion of MYR1, MYR2, and MYR3 results in 237 a small but significant, negative effect on parasite growth in vitro (11). To determine if 238 disruption of the three new genes involved in effector translocation described here has a monolayers 7 days post infection, and measured plaque size. The results show that the 241 ∆myr4, ∆gra44 and ∆gra45 strains all exhibit a significant growth defect compared to the 242 parental strain (Fig. 3F). We did not test for rescue of the growth phenotype with 243 complementation due to limitations in selectable markers available in these strains. The 244 ∆gra54 and ∆ppm3c strains, on the other hand, did not have significant growth defects, 245 consistent with the growth defects observed being dependent on the respective 246 genotype rather than nonspecific effects of the manipulations. 247 To confirm that ablation of GRA44, GRA45, and MYR4 loci are responsible for 248 the observed defect in GRA16 export, we transiently expressed a C-terminally V5-  Interestingly, GRA44, GRA45 and MYR4 all contain one or two instances of the 259 three-amino-acid motif "RRL" (Fig. 5A), which has previously been shown to be the 260 preferred sequence for cleavage by ASP5 protease (15). Indeed, cleavage at the three 261 sites shown in GRA44 and GRA45 (27), as well as at the first "RRL" motif in the essential for the translocation of all GRA effectors so far tested (5, 6, 8, 9, 15-17) and it 264 has previously been suggested that ASP5-mediated cleavage of some effectors is 265 required to "license" them for translocation across the PVM, as appears to be the case 266 in Plasmodium (38, 39). Given, however, that not all such effectors contain ASP5 267 processing motifs (e.g., GRA24 lacks the canonical "RRL" and shows no evidence of 268 ASP5-dependent processing; (17)), and given that the three newly identified 269 components of the translocation machinery identified here do, we hypothesized that 270 ASP5's essential contribution to effector translocation across the PVM might be in 271 processing one or more components of the translocation machinery. We have 272 previously shown that MYR1 is also processed by ASP5 at a "RRL" site but this does 273 not appear to be required for MYR1 to function in effector translocation (24) and so we 274 turned our attention to the newly identified translocation components identified here. 275 To determine if processing at the "RRL" sites of GRA44, GRA45, and MYR4 is 276 required for protein translocation activity, we mutated the ASP5 cleavage sites by 277 converting the first arginine to an alanine (i.e., RRLàARL) in the V5-tagged 278 complementation plasmids for each gene, and transiently transfected these into the 279 corresponding disrupted line. Western blots were then used to show that processing of 280 GRA45 at its lone "RRL" and of GRA44 at its second "RRL" is indeed abrogated by the 281 mutations (Fig. 5B). For the more N-terminal site in GRA44 (R83A), we cannot 282 definitively confirm that the mutation abrogates ASP5 processing because GRA44 is 283 epitope-tagged at its C-terminus and so, assuming cleavage at the two sites is an 284 independent event, cleavage at the downstream site will produce a C-terminal, V5-  Interestingly, mutation of the "RRL" to an "ARL" in MYR4 did not appear to affect 291 the processing of the protein (Fig. 5B). To rule out whether this is due to incomplete 292 ablation of the ASP5 processing site with a single amino acid substitution, we assessed 293 the processing of an RRLàAAA MYR4 mutant where the entire ASP5-processing motif 294 is mutated to alanines. The results (Fig. 5C) show that the higher molecular weight 295 product of MYR4 (~130kD) does not change in mobility upon mutation of the entire 296 "RRL" motif, and thus we conclude that little if any MYR4 is processed by ASP5. Note 297 that, despite repeated attempts with large amounts of DNA, the signal for the transiently 298 expressed MYR4 was never strong enough to confidently conclude whether a small 299 amount of a processed form might be present in these transiently transfected parasites; 300 we therefore cannot comment on whether the low-intensity, smaller molecular weight 301 product of MYR4 (~100kD) seen in long exposures of endogenously tagged wild type 302 MYR4 (Figs. 2A, S1A) is a result of an ASP5 processing event. 303 Having generated the four RRLàARL mutants, and having validated that ASP5 304 cleavage is ablated in at least two instances, we next tested each for its impact on the 305 localization of the epitope-tagged, C-terminal portion of the protein and on the ability of 306 the uncleaved protein to function; i.e., whether it can rescue the defect in effector 307 protein translocation. The results show that the RRLàARL mutated versions of each 308 protein are still secreted into the PV, similar to the wild type copy (Figs. 6A, 4A), and 309 are all able to rescue the translocation defect to a similar extent as the corresponding 310 control (WT) plasmid (Fig. 6B). While the GRA45 R64A mutant did substantially rescue 311 translocation, it did not consistently rescue to wildtype levels. Nevertheless, these data 312 suggest that mutation of the "RRL" sites in GRA44, GRA45, and MYR4 to "ARL" does 313 not substantially affect their function in effector protein translocation. 314 Using affinity purification of MYR1 under conditions expected to retain 316 associating partners, we identify three novel parasite proteins, GRA44, GRA45, and 317 MYR4, as essential for the export of GRA effectors into infected cells. Additionally, we 318 localize MYR4, as well as two additional MYR1-associating proteins, GRA54 and 319 PPM3C, to the PV in infected cells. Altogether, eight proteins are now known to be 320 necessary for effector export: the 3 described here and MYR1, MYR2, MYR3, 321 ROP17 and ASP5 -summarized in Table S1 ( and Neospora caninum (Table S1). 332 GRA44, by contrast, contains a putative phosphatase domain that shares 333 homology to a region of the Plasmodium serine/threonine phosphatase UIS2 (28% 334 identity over 21% of the protein; BLASTP 2.10.0+), which has recently been shown to 335 localize to the Plasmodium PVM in liver stage parasites (40). Whether UIS2 plays a 336 role in protein translocation in Plasmodium remains to be determined but this would be surprising given that none of the other components of the complex known to promote 338 translocation in Plasmodium (known as PTEX) so far studied play a role in translocation 339 in Toxoplasma (2). Additionally, whether this phosphatase domain is important for 340 effector export in Toxoplasma is not yet known. Given that the kinase domain of ROP17 341 is necessary for GRA16 export (12) it is intriguing that two of the eight factors necessary 342 for effector export are either a kinase or a phosphatase. There are numerous serine 343 residues that are phosphorylated among MYR1, MYR2, MYR3, and MYR4, supporting 344 the possibility that phosphorylation of the translocation machinery is critical to regulating 345 its function in effector export. While this work was in progress, we learned of similar 346 studies by Blakely, Arrizabalaga and colleagues who also found that GRA44 associates 347 with MYR1 and is necessary for efficient c-Myc upregulation during infection (see 348 accompanying manuscript). These latter authors used a knockdown approach to study 349 GRA44 and saw a more dramatic impact of GRA44 loss on parasite growth than we 350 report here for the GRA44 knockout; this might indicate that compensatory changes 351 were selected for during the prolonged selection necessary to generate and expand our 352 knockout clone, as was reported for AMA1 knockouts that showed dramatic up-353 regulation of the paralogue, AMA2 (41). Thus, transcriptomic analysis of the GRA44 354 knockout may reveal clues to its specific role(s) in Toxoplasma tachyzoites. 355 Our results expand the enigmas of why some parasite proteins are proteolytically 356 processed by ASP5, and why ASP5 is essential for effector translocation across the 357 PVM. MYR1, GRA44, and GRA45 all possess "RRL" motifs that appear to be cleaved in 358 an ASP5-dependent manner yet, surprisingly, their function in the export of GRA16, and 359 of GRA24 in the case of MYR1 (24), appears agnostic to mutation of these sites. 360 stay connected through a disulfide bond after cleavage (11); it remains to be determined 362 whether the polypeptides formed by RRL cleavage in GRA44 and GRA45 likewise 363 associate in a similar manner. It is also important to note that our assays may not be 364 sensitive enough to detect small changes in protein abundance in the host nucleus, and 365 that it is the combination of multiple proteins not being processed by ASP5 that is 366 deleterious to export in Dasp5 mutants, rather than the result of failure to cleave any  Lastly, none of GRA44, GRA45, or MYR4 were identified in the forward genetic 381 screen of parasites that are unable to induce c-Myc (10). This could be due to the 382 growth defects observed in ∆myr4, ∆gra44, and ∆gra45 parasites shown here since 20 parasites with null mutations in these genes might be lost during the 7-8 rounds of 384 selection used in that screen due to a fitness disadvantage. Alternatively, the 385 mutagenesis-based genetic screen was not saturating and so a more comprehensive, 386 genome-wide screen using CRISPR/Cas9 technologies might reveal these and other 387 genes responsible for effector translocation in Toxoplasma. Regardless, our finding of 388 three new components of the export machinery provides a richer understanding of how 389 Toxoplasma delivers effectors into host cells. Future work will determine the precise 390 function of each, including how they interact, the role of ASP5 cleavage, and which, if 391 any, constitutes the actual translocon. 392

Parasite strains, culture and infections 397
All Toxoplasma tachyzoites used in this study are in the Type I "RH" background, and incubated with secondary antibodies for 1 hour at RT. Vectashield with DAPI stain 426 (Vector Laboratories) was used to mount the coverslips on slides. Fluorescence was 427 detected using wide-field epifluorescence microscopy and images were analyzed using 428 ImageJ. All images shown for any given condition/staining in any given 429 comparison/dataset were obtained using identical parameters.

Endogenous tagging 460
Endogenous tagging plasmids were transfected into Toxoplasma via 461 electroporation. Tachyzoites were allowed to infect HFFs in T25 flasks for 24 hours, 462 after which the medium was changed to complete DMEM supplemented with 50 μg/ml guanine-phosphoribosyltransferae (HXGPRT or HPT) marker for 3-5 days. 465

Gene disruption 466
A list of all sgRNA sequences used in this study can be found in File S2. 467

Ectopic expression 479
Plasmids for ectopic expression were transiently transfected into Toxoplasma 480 using electroporation. Tachyzoites were allowed to infect HFFs for 18-24 hours before 481 assessing for expression of the ectopically expressed protein via either IFA or western 482 blotting. 483

Western blotting 484
blocked with 5% nonfat dry milk in TBS supplemented with 0.5% Tween-20, and 488 proteins were detected by incubation with primary antibodies diluted in blocking buffer 489 followed by incubation with secondary antibodies (raised in goat against the appropriate 490 species) conjugated to horseradish peroxidase (HRP) and diluted in blocking buffer. HA 491 was detected using a horseradish peroxidase (HRP)-conjugated HA antibody (Roche 492 cat no. 12013819001), SAG2A was detected using rabbit polyclonal anti-SAG2A 493 antibodies (48), Cyclin E1 was detected using mouse monoclonal antibody HE12 (Santa 494 Cruz Biotechnology), and GAPDH was detected using mouse monoclonal anti-GAPDH 495 antibody 6C5 (Calbiochem). Horseradish peroxidase (HRP) was detected using 496 enhanced chemiluminescence (ECL) kit (Pierce). 497

Plaque assay 498
Parasites were syringe-released from HFFs and added to confluent HFFs in T25 499 flasks. After 7 days, the infected monolayers were washed with PBS, fixed with 500 methanol, and stained with crystal violet. Plaque area was measured using ImageJ. 501

Immunoprecipitations (IPs) for mass spectrometry 502
IPs to identify MYR1-interacting proteins in HFFs were performed as follows. 503 One 15-cm dish of HFFs for each infection condition was grown to confluence. HFFs 504 were infected with either 15 x 10 6 RH::MYR1-3xHA or RH∆hpt parasites for 24 hours. 150V. The gel was washed once in UltraPure water (Thermo), fixed in 50% methanol 523 and 7% acetic acid for 15 min, followed by 3 additional washes with UltraPure water. 524 The gel was stained for 10 min with GelCode Blue (Thermo) and washed with UltraPure 525 water for an additional 20 min. One gel band (approx. 1.5 cm in size) for each condition 526 was excised and de-stained for 2 hours in a 50% methanol and 10% acetic acid 527 solution, followed by a 30 min soak in UltraPure water. Each gel slice was cut into 1 mm 528 x 1 mm squares, covered in 1% acetic acid solution, and stored at 4 °C until the in-gel 529 digestion could be performed. 530 To prepare samples for mass spectrometry, the 1% acetic acid solution was 531 removed, 10 µl of 50 mM DTT was added, and volume was increased to 100 µl with 50 mM ammonium bicarbonate. Samples were incubated at 55 °C for 30 min. Samples 533 were then brought down to RT, DTT solution was removed, 10 µl of 100 mM acrylamide 534 (propionamide) was added and volume was again normalized to 100 µl with 50 mM 535 ammonium bicarbonate followed by an incubation at RT for 30 min. Acrylamide solution 536 was removed, 10 µl (0.125 µg) of Trypsin/LysC (Promega) solution was added and 537 another 50 µl of 50 mM ammonium bicarbonate was added to cover the gel pieces. 538 Samples were incubated overnight at 37 °C for peptide digestion. Solution consisting of 539 digested peptides was collected in fresh Eppendorf tubes, 50 µl of extraction buffer 540 (70% acetonitrile, 29% water, 1% formic acid) was added to gel pieces, incubated at 37 541 °C for 10 min, centrifuged at 10,000 x g for 2 minutes and collected in the same tubes 542 consisting of previous elute. This extraction was repeated one more time. Collected 543 extracted peptides were dried to completion in a speed vac and stored at 4 °C until 544 ready for mass spectrometry. 545

Mass spectrometry 546
Eluted, dried peptides were dissolved in 12.5 μl of 2% acetonitrile and 0.1% 547 formic acid and 3 μl was injected into an in-house packed C18 reversed phase 548 analytical column (15 cm in length). Peptides were separated using a Waters M-Class 549 UPLC, operated at 450 nL/min using a linear 80 minute gradient from 4-40% mobile 550 phase B. Mobile phase A consisted of 0.2% formic acid, 99.8% water, Mobile phase B 551 was 0.2% formic acid, 99.8% acetonitrile. Ions were detected using an Orbitrap Fusion 552 mass spectrometer operating in a data dependent fashion using typical "top speed" 553 methodologies. Ions were selected for fragmentation based on the most intense multiply 554 charged precursor ions using Collision induced dissociation (CID). Data from these 555 analyses was then transferred for analysis. 556

Mass spectrometric analysis 557
The .RAW data were searched using MaxQuant version 1.6.1.0 against the 558 canonical human database from UniProt, Toxoplasma GT1 databases from ToxoDB 559 parasites. The populations of endogenously tagged parasites analyzed in Fig. 2A were 822 seeded onto empty coverslips before being fixed with methanol. The corresponding 823 tagged proteins were detected with rat anti-HA antibodies, the marker for dense granule 824 proteins, GRA7, was detected with rabbit anti-GRA7 antibodies, and the parasites were 825 visualized with differential interference microscopy (DIC). Scale bar is 5µm. uninfected HFF lysate, for 20 hours before lysates were generated for immunoblotting. 853 Lysates were analyzed by western blotting using mouse anti-cyclin E1 antibodies. 854 Rabbit anti-SAG2A was used to assess the levels of parasite protein in the lysate. 855 Lyse cells, sonicate, and add lysate to anti-HA beads