Toxoplasma gondii Toxolysin 4 Contributes to Efficient Parasite Egress from Host Cells

ABSTRACT Egress from host cells is an essential step in the lytic cycle of T. gondii and other apicomplexan parasites; however, only a few parasite secretory proteins are known to affect this process. The putative metalloproteinase toxolysin 4 (TLN4) was previously shown to be an extensively processed microneme protein, but further characterization was impeded by the inability to genetically ablate TLN4. Here, we show that TLN4 has the structural properties of an M16 family metalloproteinase, that it possesses proteolytic activity on a model substrate, and that genetic disruption of TLN4 reduces the efficiency of egress from host cells. Complementation of the knockout strain with the TLN4 coding sequence significantly restored egress competency, affirming that the phenotype of the Δtln4 parasite was due to the absence of TLN4. This work identifies TLN4 as the first metalloproteinase and the second microneme protein to function in T. gondii egress. The study also lays a foundation for future mechanistic studies defining the precise role of TLN4 in parasite exit from host cells. IMPORTANCE After replicating within infected host cells, the single-celled parasite Toxoplasma gondii must rupture out of such cells in a process termed egress. Although it is known that T. gondii egress is an active event that involves disruption of host-derived membranes surrounding the parasite, very few proteins that are released by the parasite are known to facilitate egress. In this study, we identify a parasite secretory protease that is necessary for efficient and timely egress, laying the foundation for understanding precisely how this protease facilitates T. gondii exit from host cells.

Such studies have solidified roles for cyclic GMP (cGMP) and calcium to stimulate protein kinase G (PKG) (4,5) and calcium-dependent protein kinases (CDPKs) (4,(6)(7)(8), among other targets. Calcium signaling results in the activation of parasite motility and the discharge of apical secretory granules, termed micronemes (9)(10)(11)(12). While the importance of microneme proteins, such as transmembrane adhesins that connect with the actin-myosin motor system to drive gliding motility and active cell invasion, have been established, microneme proteins also contribute to active egress via their role in the formation of pores such as those created by Perforin-like protein 1 (PLP1) (13)(14)(15). Gene knockout (KO) studies suggest that PLP1 pore formation disrupts the parasitophorous vacuole membrane (PVM), which encases parasites during intracellular replication. PLP1-deficient parasites are delayed or fail in egress and show a marked loss of virulence in infected mice, implying a link between efficient egress and virulence. Although several proteins released from parasite-dense granules (calcium-independent secretory organelles released during parasite replication) have also been implicated in egress (16)(17)(18)(19)(20), PLP1 is the only microneme protein known to directly function in egress to date.
Previous work identified a putative metalloprotease, toxolysin 4 (TLN4; TGME49_ 206510), in a proteomic screen of Toxoplasma secretory products released by extracellular tachyzoites (21). A subsequent study showed that TLN4 is a microneme protein that undergoes extensive proteolytic processing and potentially contributes to parasite fitness, based on loss of TLN4-deficient parasites from a mixed population of parasites transfected with a knockout plasmid (22). However, a recent genome-wide CRISPR/Cas9 knockout screen indicated that TLN4 does not contribute substantially to parasite fitness based on its phenotype score of 0.51 (on a scale of 27 to 3, with lower values indicating lower fitness) (23). TLN4 is a member of the M16 metalloproteinase subfamily of so-called cryptases, which are represented throughout the tree of life and include four TLN genes in the Toxoplasma genome. M16 metalloproteinases are exemplified by the human insulin-degrading enzyme (IDE) and feature a catalytic chamber (or crypt) that accommodates small polypeptides for degradation. Whereas characterization of TLN2 (TGME49_227948) and TLN3 (TGME49_257010) has not been reported, TLN1 (TGME49_269885) resides in the parasite secretory rhoptries, where it plays an unknown role apart from contributing modestly to growth in vitro (24).
Here, we show that TLN4 has the structural features of an active metalloproteinase and that it possesses proteolytic activity against a model polypeptide substrate. We also report that TLN4-deficient parasites show normal gliding motility, cell invasion, and replication but have a delayed induced-egress phenotype. Our findings suggest that TLN4 contributes to Toxoplasma egress, identifying it as only the second microneme protein implicated in this event.

RESULTS
TLN4 has the structural features of an active metalloproteinase. TLN4 is a large (2,435-amino-acid [aa]) protein that contains multiple domains, including one active (A) and three inactive (IA) M16 proteinase domains, along with a long C-terminal extension that includes a repeat domain consisting of eight tandem repeats of a 28-aa sequence (Fig. 1A). Structural modeling of the M16 proteinase domains (residues 202 to 1367) using the human insulin-degrading enzyme as a template suggests that each domain forms a similar ab roll fold, with the domains arranged around a central chamber (Fig. 1B), as expected for an M16 family metalloproteinase (25). When viewed from a perspective of inside the chamber or crypt, a putative active site is visible (Fig. 1C) with the characteristic HXXEHX N EX 6 E binding motif for catalytic Zn 21 arranged on two a-helices (Fig. 1D) that are separated by 60 amino acids in TLN4 (Fig. 1E). These features are consistent with TLN4 being an active metalloproteinase of the M16 family.
Recombinant TLN4 is capable of processing b-insulin. M16 family proteinases typically act upon peptides and polypeptides that are sufficiently small to fit in the crypt. To determine if TLN4 possesses proteolytic activity, we expressed and purified a recombinant form of TLN4 lacking the C-terminal extension. TLN4 209-1295 was expressed using a bacterial expression system and was purified and refolded. The purified protein migrated on SDS-PAGE as a prominent 130-kDa band and behaved as a monodispersed protein when analyzed by size exclusion chromatography ( Fig. 2A). Upon incubating recombinant TLN4 with b-insulin as a model substrate, we observed the generation of several cleavage products (Fig. 2B), which were confirmed by mass spectrometry to originate from b-insulin (Fig. 2C). Mapping of the cleavage sites on the sequence of b-insulin revealed a preference for TLN4 cleavage in the central region of the polypeptide but no obvious preference for recognition of specific amino acids. Collectively, our findings suggest that TLN4 is an active protease that is capable of processing a model substrate.
TLN4 is amenable to genetic disruption. In a previous study of TLN4, it was proposed that TLN4 contributes to parasite fitness, because TLN4-knockout parasites could be detected in a population of transfected RH parasites but were lost through further culturing (22). To enhance the recovery of knockout parasites, we utilized RHDku80Dhxg parasites (wild type [WT] here) and transfected them with a  hypoxanthine xanthine guanine phosphoribosyl transferase (HXGPRT) selectable cassette flanked by 59 and 39 TLN4 homology regions (Fig. 3A). Individual knockout parasite clones were tested by PCR for integration of the selectable marker at the 59 and 39 ends (Fig. 3B) and for the absence of the TLN4 gene (Dtln4) (Fig. 3C). We then genetically complemented Dtln4 parasites by expression of the TNL4 cDNA containing two copies of a hemagglutinin (HA) epitope tag inserted at aa 847 after the first inactive domain (TLN4-IA 1 -HA 2 ) to generate Dtln4TLN4 parasites, which were confirmed by PCR (Fig. 3C). The difference in fragment size in the WT and Dtln4TLN4 parasites is due to the presence of introns in the WT but not in the cDNA of the complement strain. Expression of TLN4 in Dtln4TLN4 parasites was confirmed by fluorescence microscopy (Fig. 3D). We further validated expression by Western blotting, which showed the expected doublet of processed TLN4 at ;50 kDa and a smaller ;35-kDa species (Fig. 3E). This smaller species showed the expected increase in size in Dtln4TLN4 due to the HA tag. The loss of the TLN4 protein in Dtln4 was also confirmed by Western blotting (Fig. 3E).  TLN4 does not contribute to invasion, replication, gliding motility, or virulence. To assess whether TLN4 plays a role in the lytic cycle, plaque assays were performed. Compared to parental parasites, Dtln4 parasites showed significantly smaller plaques ( Fig. 4A and B), suggesting a defect in one or more steps in the parasite lytic cycle. This phenotype was fully restored in the complemented strain. We next tested each step in the lytic cycle to identify the basis of smaller plaques. Parasites deficient in TLN4 showed normal invasion (Fig. 4C) and replication at 17 h and 26 h postinfection (Fig. 4D), and all types of gliding motility were observed (Fig. 4E). Finally, no significant differences were observed in mouse survival of acute infection upon intraperitoneal infection with 10 or 100 tachyzoites (Fig. 4F).
Dtln4 parasites are defective in efficient egress. The last step of the lytic cycle is egress of the intracellular parasites from host cells, which can be induced by addition of the calcium ionophore A23187. Whereas most wild-type parasites egressed from host cells within 2 min of ionophore addition, at which time the cells were fixed, significantly fewer Dtln4 parasites egressed in the allotted time (Fig. 5A). This phenotype is mostly reversed in Dtln4TLN4 parasites. The deficiency in egress was more pronounced when zaprinast, a phosphodiesterase inhibitor, was used as an inducer (Fig. 5B), which may be due to its mode of action. Whereas A23187 elevates parasite cytosolic calcium by mobilizing it from parasite and host intracellular stores and the media, zaprinast treatment triggers the release of calcium from intracellular stores via activation of protein kinase G and inositol triphosphate signaling (4,26). In this regard, zaprinast treatment likely better mimics the physiologic release of Ca 21 from intracellular stores during natural egress. An alternative and complementary measure of parasite egress is the release of host lactate dehydrogenase (LDH) into the surrounding medium due to loss of host cell integrity. After normalizing LDH release to WT parasites, Dtln4 parasites showed ;45% release of LDH, while the complement strain nearly restored LDH release to wild-type levels (Fig. 5C). In examining the excretedsecreted antigen fraction of extracellular parasites, conducted in buffers with neutral or acidic pH (to mimic PV acidification [13][14][15]), there was no change in the processing of PLP1 in the Dtln4 parasites compared to WT parasites (Fig. 5D). Together, these findings suggest a role for TLN4 in T. gondii egress from host cells that is independent of PLP1 processing.

DISCUSSION
TLN4 was previously shown to be an extensively processed microneme protein (22), but its contribution to the lytic cycle was not determined due to the inability to generate a knockout of this gene in the wild-type RH strain background. In this study, we were able to isolate a TLN4 knockout strain by utilizing the more genetically tractable RHDku80 strain. Plaque assays showed that Dtln4 parasites affected the lytic cycle, based on the smaller plaque sizes observed. Assessing the known steps of the lytic cycle showed that there were no defects in invasion, replication, or types of gliding motility performed by the parasites, and there was no effect on parasite virulence. The only deficiency observed was in induced egress, assessed by a static egress assay, and LDH release. Complementation of Dtln4 with a TLN4 construct under the control of the endogenous TLN4 promoter fully restored the plaque defect and largely restored the egress deficiency.
In vitro investigation into TLN4 suggests that the protein is an active protease based on cleavage of b-insulin as a model substrate, supporting a potential proteolytic function within the parasite. Homology modeling indicates that TLN4 adopts the typical M16 structure and contains a catalytic chamber or "crypt," which is one of the defining features of the M16 family of metalloproteinases. Studies with other M16 enzymes indicate that the crypt encapsulates and cleaves polypeptides, with the substrate being determined by the size and charge of the crypt as well as the flexibility of the substrate (s) (27,28). Although it is possible that TLN4 plays a direct role in egress by, e.g., facilitating the disruption of the PVM, an indirect contribution of TLN4 to egress is also plausible. An indirect role could involve proteolytically activating another parasite protein that contributes to egress or degrading a protein that would otherwise compromise egress. Identifying substrates of TLN4 in future studies might reveal the basis for its contribution to parasite exit from host cells.
PLP1 is the only microneme protein identified to date to have a direct role in egress. Interestingly, mice infected with Dtln4 have no virulence defect, whereas Dplp1 parasites are markedly virulence attenuated. This notable distinction could be because the egress defect of Dtln4 parasites is less pronounced than that for Dplp1, or it could be due to additional roles for PLP1 during infection of mice. Future studies identifying other secretory proteins that contribute to egress to various degrees along with other work defining how PLP1 shapes the outcome of infection should help distinguish between these possibilities.
Among the Apicomplexa, the M16A subfamily is constrained to coccidian parasites, including close (e.g., Hammondia hammondia and Neospora caninum) and more distant (Sarcocystis neurona and Eimeria spp.) relatives of T. gondii (https://orthomcl.org/ orthomcl/app). Whereas the genome of T. gondii encodes 11 M16 metalloproteinases, of which 4 belong to the M16A subfamily (29), the genome of the intestinal parasite Cryptosporidium parvum encodes an expanded family of 22 M16 metalloproteinases, 18 of which are M16A members. Among these, INS1 was recently shown to be necessary for formation of macrogamonts, the female sexual stage of C. parvum (30). The results of reciprocal BLAST searches indicate that TLN4 is most closely related to C. parvum INS-15 (cgd3_4260) and INS-16 (cgd3_4270), which are closely related orthologs (83% identical) (30,31). INS-15 localizes to the mid-apical region of C. parvum sporozoites and possibly merozoites (31). Antibodies to INS-15 impaired sporozoite cell invasion, suggesting a role in parasite entry into host cells. Future studies involving the targeted deletion of M16A family members in C. parvum and other coccidian parasites will be necessary to appreciate further their contributions to parasite infection biology.

MATERIALS AND METHODS
Structural modeling of TLN4. The full amino acid sequence of ME49 strain TLN4 was used to query Phyre2, which identified human insulin-degrading enzyme (PDB entry 2JBU) as the top-scoring template. This analysis resulted in a structural model of TLN4 encompassing amino acids 202 to 1367. The structural model was visualized using PyMOL v2.3.2.
Expression and purification of recombinant TLN4. A region encoding domains A to IA-3 of the TLN4 cDNA from RH strain (amino acids 209 to 1295) was amplified with TLN4.625.NdeI.F (59-GGAA TTCCATATGAGAGACACGAGCGCGTACTCGG-39) and TLN4.3885.Not.R (59-AAGGAAAAAAGCGGCCGCGCTAA GCCACCGGAGGCGCTCGAGGAAAGC-39) primers and cloned into the bacterial expression vector pET21a, introducing an in-frame C-terminal hexahistidine tag to the construct. The Escherichia coli expression construct containing the TLN4 209-1295 open reading frame (ORF) was verified by DNA sequencing.
The TLN4 209-1295 expression plasmid was transformed into Escherichia coli BL21 Gold cells selected on 2Â yeast-tryptone (2YT) broth-agar plates containing 100 mg/ml ampicillin. Twenty milliliters of 2YT supplemented with ampicillin was grown to confluence overnight and then added to 1 liter of 2YT containing ampicillin and incubated with shaking at 378C until an optical density at 600 nm (OD 600 ) of ;0.85 was reached. Cultures were then induced with a final concentration of 1 mM isopropyl-b-D-thiogalactopyranoside (IPTG) overnight at 168C.
Harvest and cell lysis of the overnight expression culture showed that TLN4 209-1295 expressed to high yields but was insoluble. Inclusion bodies were prepared by harvest of the insoluble fraction of the whole-cell lysate and resuspension with 30 ml of wash buffer I (25 mM Tris, pH 8.0, 300 mM NaCl, 1% Triton X-100). The suspension was homogenized until an even consistency was obtained. The inclusion bodies were then washed twice by centrifugation and homogenization in wash buffer I. Three subsequent washes with wash buffer II (25 mM Tris, pH 8.0, 300 mM NaCl) were performed to remove the excess Triton X-100. Purified inclusion bodies were stored at 2808C until required. Proteins were resolubilized in 25 mM Tris, pH 8.0, 250 mM NaCl, 6 M guanidine hydrochloride, 2 mM beta-mercaptoethanol by rotation at 48C before centrifugation. Harvest supernatant was applied to nickel-nitrilotriacetic acid resin and purified via metal affinity chromatography. Purity of the denatured purified protein was assessed via SDS-PAGE and Western blotting.
Denatured TLN4 209-1295 was diluted to a final concentration of ,1 mg/ml and applied dropwise to 100 ml of refold buffer (50 mM Tris, pH 8.2, 250 mM NaCl, 0.5 M arginine, 0.44 M sucrose, 4 mM reduced glutathione, 0.4 mM oxidized glutathione, 0.05 mM Tween 20) that was constantly stirring at 48C. Refolding proceeded for 4 h at 48C with gentle stirring. The refolded protein was then dialyzed overnight against 25 mM Tris, pH 8.0, 300 mM NaCl at 48C. Postdialysis, the protein was concentrated to approximately 0.5 ml before application to a Superdex 200 10/300 column for further purification via size-exclusion chromatography.
TLN4 proteolytic activity. Hydrolysis of insulin B chain was measured via reverse-phase high-performance liquid chromatography (HPLC) following incubation of TLN4 (160 mg/ml) with insulin B chain (42 mM; Sigma-Aldrich) in 10 mM Tris for 88 h. HPLC was carried out in a Vydac C 4 HPLC column using a linear gradient from 0.1% trifluoroacetic acid (TFA) in 95% water, 5% acetonitrile to 0.1% TFA in 50% acetonitrile-water at a flow rate of 1 ml/min. Hydrolysis products were detected at 214 nm and collected for analysis. The obtained peptide products were analyzed on an Applied Biosystems 4800 matrix-assisted laser desorption ionization tandem time-of-flight proteomics analyzer at the University of Kentucky Proteomics Core.
The TLN4-IA 1 -HA 2 complementation plasmid was previously described in reference 22. This plasmid (50 mg) was cotransfected with 5 mg of the pDHFR-TS plasmid containing the selectable marker dihydrofolate reductase-thymidylate synthase by electroporation and drug selection with pyrimethamine. Positive clones were identified by PCR and Western blot analysis.
Invasion assays. Invasion assays were performed as previously described (32). Briefly, 1 Â 10 7 parasites were used to infect subconfluent HFF monolayers in 8-well chamber slides for 20 min before fixation with paraformaldehyde. Slides were differentially stained with anti-SAG1 antibodies to differentiate attached versus invaded parasites.
Egress assays. Induced egress was performed as described previously (15). Briefly, parasites grown in HFFs in an 8-well chamber slide for 30 h were treated with 1% dimethyl sulfoxide, 2 mM A23187, or 200 mM zaprinast in assay buffer (Hanks' buffered salt solution containing 1 mM CaCl 2 , 1 mM MgCl 2 , and 10 mM HEPES) and incubated for 2 min in a 37°C water bath. Egress was stopped by addition of 2Â fixative (8% formaldehyde in 1Â PBS). Immunofluorescence was performed with rabbit anti-SAG1 to identify parasites and mouse anti-GRA7 (a generous gift from Peter Bradley) to identify the parasitophorous vacuole membrane. At least 10 fields of view (400Â total magnification) per condition were enumerated as occupied or unoccupied.
LDH egress assays were performed as previously described (20). Briefly, parasites were grown in HFF monolayers in 96-well plates for 30 h. Wells were washed with Ringer's buffer and then treated with 100 mM zaprinast diluted in Ringer's buffer. Plates were incubated at 378C for 20 min before removal and placement on ice. Fifty microliters of the supernatant were removed and centrifuged in a separate round-bottom plate. Thirty microliters of supernatant was subsequently removed and release of lactate dehydrogenase was determined using an LDH cytotoxicity colorimetric assay kit (BioVision).
Plaque and replication. Parasites were inoculated into wells of a 6-well plate and allowed to replicate undisturbed for 7 days. The wells were then stained with 0.2% crystal violet for 5 min and rinsed with double-distilled water (ddH 2 O). Images of the wells were scanned, and plaque number and size were analyzed with Image J.
For the replication assay, cells of an 8-well chamber slide were inoculated with 1.25 Â 10 5 tachyzoites and allowed to invade and grow for 17 h and 26 h prior to fixation and indirect immunofluorescence.
Live gliding video microscopy. Glass-well dishes (MatTek) were coated with 50% fetal bovine serum (FBS)-50% PBS overnight at 48C or 1 h at 378C and washed with PBS before use. Parasites were filter purified and resuspended in HHE (Hanks' salt solution, 10 mM HEPES, EDTA). Parasites were then added to coated dishes and allowed to settle for 5 min at room temperature. Dishes were then moved to a 378C chamber with 5% CO 2 and warmed for 5 min prior to starting video recording. Each video consisted of 1 frame/s recorded for 90 s. Enumeration of the types of gliding motility were carried out by examining the videos in addition to maximum projection images generated by the Zeiss AxioVision software.
Mouse infections. All laboratory animal work in this study was carried out in accordance with policies and guidelines specified by the Office of Laboratory Animal Welfare, the U.S. Department of Agriculture, and the American Association for Accreditation of Laboratory Animal Care (AAALAC). The University of Michigan Committee on the Use and Care of Animals (IACUC) approved the animal protocol used for this study (Animal Welfare Assurance A3114-01, protocol no. PRO00008638). Freshly egressed parasites were filter purified in PBS, washed, counted, and injected intraperitoneally in 200ml of PBS into 6-to 8-week-old female Swiss Webster mice. The experiment was performed once with 4 groups of 12 mice each infected with 10 or 100 tachyzoites of WT or Dtln4 parasites. Mice were given water and food ad libitum, monitored twice daily, and were humanely euthanized upon showing signs of morbidity.
Western blotting. Parasite lysates were generated by filter purifying parasites followed by centrifugation, 1Â wash with cold PBS, and resuspension in .908C 1Â sample buffer. Lysates were boiled for 5 min prior to running on SDS-PAGE gels. Gels were semidry electroblotted (Bio-Rad) onto polyvinylidene fluoride membranes and sequentially probed with mouse anti-TLN4 (22) and horseradish peroxidase-conjugated goat anti-mouse (Jackson ImmunoResearch). Bands were revealed by enhanced chemiluminescence with SuperSignal West Pico Plus chemiluminescent substrate (Thermo Fisher Scientific) and documented with a Syngene Pxi imaging system. PLP1 processing. Extracellular parasites (Dku80Dhxg [WT], Dplp1, Dtln4, and Dtln4TLN4) were resuspended in PBS in either pH 5.4 or 7.4, and excreted-secreted antigen (ESA) was collected. Pellet lysates and ESA were separated on SDS-PAGE gels, and membranes were probed with antibodies against rabbit anti-PLP1.