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.

Keywords: Toxoplasma gondii; apicomplexan parasites; egress; intracellular parasites; proteases.

FIG 1

Structural features of TLN4. (A) Domain structure of TLN4 illustrating its possession of a signal peptide (SP) and, based on the IDE structure, a single active domain (domain A), 3 inactive domains (IA^1-3), and a C-terminal repeat domain. Note that a largely featureless region between IA³ and the repeat domain is not shown and is instead denoted by //. Numbering indicates amino acid positions and is based on the sequenced cDNA of TLN4 (22). (B) Structural model of TLN4 showing the arrangement of domains into a shell surrounding a central catalytic chamber. Three different views are provided. (C) The A and AI³ domains viewed from the center of the chamber with resides that coordinate zinc binding shown in orange. (D) Zoom-in of the adjacent alpha helices of the A domain showing the residues (orange) that are predicted to coordinate binding to zinc. (E) Sequence of the alpha helices shown in panel D, including residues (orange) that likely mediate binding to zinc as a cofactor for proteolytic activity.

FIG 2

Recombinant TLN4 proteolytic activity using β-insulin as a model substrate. (A) Chromatogram showing the elution of recombinant TLN4 as a monodispersed peak from size exclusion chromatography. SDS-PAGE of purified TLN4 is shown inset on the left, with molecular weight markers indicated on the right of the elution peak. Values in parentheses indicate molecular size in kDa. (B) HPLC profile of insulin B chain cleaved by TLN4. Hydrolysis products were determined by comparing a reaction allowed to incubate (top line) versus a reaction that was immediately stopped (bottom line). (Inset) Captured peaks are labeled with their residue number, and a cleavage map was developed. (C) Mass spectrometry analysis of captured hydrolysis products from insulin B chain are shown with their expected mass and observed mass along with the fragment and cleavage sites that each product represents.

FIG 3

Genetic disruption of TLN4. (A) Schematic showing the TLN4 genomic locus and homologous recombination with a cassette containing a TLN4 5′ flank, the HXGPRT selectable marker, and a TLN4 3′ flank. Arrows and letters indicate the placement of primers used in panel B, and the black bar in TLN4 indicates an intron that permits distinction of the gene from the cDNA in panel C. Other introns of TLN4 are not shown. (B) PCR validation of selectable marker integration at the 5′ (primers a and b) and 3′ (primers c and d) end. M, molecular weight marker. (C) PCR validation of TLN4 deletion (Δtln4) and complementation (Δtln4TLN4) with the cDNA (primers e and f). The differences in size are due to the presence of an intron in the WT genomic DNA versus the cDNA in the complement strain lacking the intron. (D) Schematic showing the complementation construct with a 2× HA tag in the full-length TLN4 coding sequence. The immunofluorescence panel shows the micronemal localization of the complemented strain, costained with anti-MIC2. (E) Mouse antibodies to TLN4 detect the ∼55-kDa doublet and the ∼32 kDa of TLN4 in the lysates of WT and Δtln4TLN4 parasites. Slower mobility of the smallest band in the complement strain is due to the HA tag. Sizes (in kilodaltons) of molecular weight markers are shown to the left. Detection of T. gondii tubulin was included as a loading control.

FIG 4

Effect of TLN4-deficient parasites on the lytic cycle. (A) Plaque assays show smaller plaques in Δtln4 parasites. One hundred parasites of each strain were inoculated into 6-well plates and allowed to grow for 7 days undisturbed. Wells were stained with crystal violet. (B) Individual plaque sizes were measured using ImageJ. Data represent three independent biological experiments with triplicate samples within each experiment. A minimum of 500 plaques were counted per strain. Error bars represent standard errors of the means (SEM), and statistics were performed using an unpaired, two-tailed, Student's t test. (C) Parasites lacking TLN4 show normal invasion. HFF cells in 8-well chamber slides were inoculated with parental, Δtln4, or Δtln4TLN4 parasites and allowed to invade for 20 min prior to fixation. Wells were differentially stained with SAG1 antibodies to detect attached or invaded parasites per host cell nucleus. A minimum of 250 host cells were counted per strain. Data in the invasion (C) and replication (D) graphs represent means ± SEM from three independent experiments, each with triplicate samples. At least 8 fields of view were counted per well. (D) Parasites lacking TLN4 replicate normally. Parasites were inoculated into 8-well chamber slides and allowed to replicate for 17 h or 26 h prior to fixation and enumeration of parasites per vacuole. A minimum of 250 vacuoles were counted per strain per time point. (E) Δtln4 parasites show normal modes of gliding motility. Video microscopy to enumerate the types of gliding motility shows the percentages of immobile parasites and parasites performing twirling, circular, and helical gliding in all strains tested. Graphs indicate the means ± SEM from three independent experiments. *, P ≤ 0.05 by Student's t test. (F) TLN4 does not play a role in acute virulence. Swiss-Webster mice were infected intraperitoneally with 10 or 100 tachyzoites of WT or Δtln4 parasites, and survival time was enumerated; 12 mice were infected for each strain and inoculum.

FIG 5

TLN4-deficient parasites are defective in egress. (A and B) Parasite egress was quantified by immunofluorescence microscopy. Thirty-hour vacuoles were treated with DMSO or 2 μM A23187 (A) or 200 μM zaprinast (B) for 5 min prior to fixation and enumeration. *, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.001; ****, P ≤ 0.0001; all by Student's t test. A minimum of 500 vacuoles were counted per strain. (C) Lactate dehydrogenase (LDH) release following induction with 100 μM zaprinast was used as a measure of egress. Data are normalized to WT LDH release. All graphs shown are means ± SEM from three biological replicates, each with triplicate samples. (D) Parasites lacking TLN4 have normal PLP1 processing. Pellet and ESA were collected from parasites at pH 5.4 and 7.4, separated on SDS-PAGE, blotted onto membranes, and probed with rabbit anti-PLP1 antibodies. Processing of PLP1 in the ESA of all strains expressing PLP1 observed at pH 5.4. The pellet blot (bottom) acts as a loading control.