A plasma membrane localized protein phosphatase in Toxoplasma gondii, PPM5C, regulates attachment to host cells

Abstract

The propagation of Toxoplasma gondii is accomplished by repeated lytic cycles of parasite attachment to a host cell, invasion, replication within a parasitophorous vacuole, and egress from the cell. This lytic cycle is delicately regulated by calcium-dependent reversible phosphorylation of the molecular machinery that drives invasion and egress. While much progress has been made elucidating the protein kinases and substrates central to parasite propagation, little is known about the relevant protein phosphatases. In this study, we focused on the five protein phosphatases that are predicted to be membrane-associated either integrally or peripherally. We have determined that of these only PPM5C, a PP2C family member, localizes to the plasma membrane of Toxoplasma. Disruption of PPM5C results in a slow propagation phenotype in tissue culture. Interestingly, parasites lacking PPM5C divide and undergo egress at a normal rate, but have a deficiency in attaching to host cells. Both membrane localization and phosphatase activity are required for PPM5C's role in attachment. Phosphoproteomic analysis show relatively few phosphorylation sites being affected by PPM5C deletion in extracellular parasites of which several are found on proteins involved in signaling cascades. This implies that PPM5C is part of a wider regulatory network important for attachment to host cells.

Figure 1

Schematic of five putative membrane associated protein phosphatases. Predicted functional domains and protein modification sites are shown for PPM2A (TGGT1_232340), PPM2B (TGGT1_267100), PPM3D (TGGT1_202610), PPM5C (TGGT1_281580), PPM11C (TGGT1_304955). Below PPM11C are the two putative proteins encoded genomic locus of TGGT1_304955. M, myristoylation; P, palmitoylation; SP, signal peptide; TM, transmembrane domain; PP2Cc, PP2C phosphatase catalytic domain.

Figure 2

Localization of the protein phosphatases. (A) Intracellular parasites expressing endogenously HA tagged protein phosphatases were stained with antibodies against the HA epitope (red) and the surface protein Sag1 (green). Panels to the right show western blot of parasite extracts probed for HA. Original blots are shown in Supplemental Fig. S1. Arrow in PPM5C row points at PPM5C in the residual body. Scale bar = 2 µm. (B) Intracellular parasites were stained with antibodies against the HA epitope and IMC3, a component of the Inner Membrane Complex. Scale bar: 5 µm.

Figure 3

Role of amino acids predicted to be myristoylated or palmitoylated in PPM5C localization. Parasites were transfected with plasmids encoding either wild type or mutant PPM5C to determine the role of glycine 2, which is predicted to be myristoylated and cysteine 4, predicted to be palmitoylated, in localization. (A) Diagram of the template PPM5C expression plasmid is shown. PPM5C coding sequences are under the control of the high expressing tubulin promoter (tub pr) and include three copies of the HA epitope tag (3xHA). Plasmid also contains the selectable marker HXGPRT. (B) Immunofluorescence of intracellular parasites expressing exogenous wildtype PPM5C (+tubPPM5C(HA), or mutant PPM5C in which either glycine 2 or cysteine 4 are mutated for alanine (+tubPPM5C G2A and +tubPPM5C(HA) C4A). Parasites were stained for HA to detect IMC2A (red) and the surface antigen Sag1 to detect the parasite membrane (green). Western blots of protein extract of each transgenic strain probed for HA are shown on the right. Original blots are shown in Supplemental Fig. S2.

Figure 4

Generation and complementation of PPM5C knockout strain. (A) Diagram depicts strategy used for developing the PPM5C knock-out strain ∆ppm5. Parasites were transformed with a plasmid encoding Cas9 and a guide RNA that targets PPM5C first exon and a repair fragment that contains regions of homology to the PPM5C locus (red and blue boxes) flanking the selectable marker DHFR. The bottom graphic in (A) represents the resulting edited genome in the knock out strain. P1 and P2 are primers used to confirm disruption. (B) Disruption of PPM5C in the knockout strain was confirmed through PCR with primers P1 and P2 shown in (A). (C) Diagram depicts plasmid used to reintroduce PPM5C into the knockout strain. In this complementation plasmid, PPM5C is driven by its own promoter (PPM5C pr) and contains a triple HA epitope tag. (D) Expression of PPM5C in the complemented strain (∆ppm5c.cp) was confirmed by immunofluorescence assays and western blot. Original blots are in Supplemental Fig. S3.

Figure 5

Phenotypic analysis of PPM5C knockout strain. Propagation, replication, invasion and attachment was determined for parasites of the parental (par), knockout (∆ppm5c) and complemented (∆ppm5c.cp) strains. (A) Parasites of all three strains were allowed to grow in human fibroblasts for six days before fixation and crystal violet staining to reveal plaques formed by repeating cycles of invasion, replication and egress. Representative images of plaque assays are shown as well as quantification of area cleared by plaques relative to parental strain. (B) Parasites were allowed to invade HFFs for 30 minutes, and infected cultures were fixed at 18, 24, and 30 hours post-infection (hpi). The proportion of vacuoles with 2, 4, 8, 16, 32 or 64 parasites was calculated for each time point and strain. (C) Same number of parasites of each strain were allowed to infect cells for 30 minutes and uninvaded parasites were washed out. After 24 hours cultures were fixed and the number of vacuoles in 20 randomly and blindly selected fields of view was quantitated. Data is presented normalized to the average of total number of vacuoles formed by the parental parasites. (D) Parasites were allowed to infect cells for 30 minutes before fixation and differential staining for parasites outside and inside of cells. Bars represent the percentage of total parasites that were inside in 10 randomly and blindly selected fields of view. (E) Same experiments as in (C) were analyzed to compare the total number of parasites both inside and outside, which represents the number of parasites that efficiently attached, in 10 randomly selected fields of view. Data is presented normalized to the average of parental parasites. For all data graphs n = 3 biological replicates × 3 experimental replicates and p-value was estimated by two tailed Student’s t-test. Error bars show standard deviations (SD). * indicates p-value < 0.05; ** indicates p-value < 0.01; *** indicates p-value < 0.001.

Figure 6

Microneme secretion in PPM5C knockout strain. Extracellular parasites of the parental (par), knockout (∆ppm5c) and complemented (∆ppm5c.cp) strains were incubated for 10 minutes with or without ethanol and spun down. Supernatant, which contains secreted antigens, and pellet, which contain whole parasites, were processed for western blots. Secreted microneme protein Mic2 and secreted dense granule protein Gra1 was detected in the supernatant (top two rows). In the absence of ethanol constitutive and baseline secretion is detected, while ethanol induces calcium dependent microneme secretion. To confirm equal number of parasites protein extract from the parasite pellet was probed with the surface antigen Sag1. Additionally, we probed for Mic2 and Gra1 in extracts from the parasite pellets to confirm equal levels of the proteins across strains tested. All original blots are shown in Supplemental Fig. S5.

Figure 7

Role of membrane localization and phosphatase activity in PPM5C function. The knockout strain ∆ppm5c was transfected with mutant versions of the complementation plasmid shown in Fig. 4C. In mutant G2A C4A, the predicted myristoylated and palmitoylated sites are changed for alanine, which mislocalize the protein to cytoplasm. In the D413A D430A mutant, two ion binding aspartate sites D413 and D430, which are required for phosphatase activity, are changed for alanine. (A) Intracellular parasites of the knockout strain complemented with either the G2A C4A or the D413A D430A mutant PPM5C were stained for HA to determine the localization exogenous mutant PPM5C. As expected, G2A C4A is mislocalized to the cytoplasm and D413A D430A correctly associates with the plasma membrane of the parasite. (B) Equal expression levels of mutant and wild type PPM5C in the complemented strains was verified by western blot probed with HA antibodies using Sag1 as a loading control. Original blots are in Supplemental Fig. S6. (C) Representative images of plaque assay of the parental (par), knockout (∆ppm5c) and mutant complemented (G2A C4A or the D413A D430A) strains are shown. Graph on the left is the quantification of the plaque area relative to parental calculated as in Fig. 5B. (D) Invasion (left) and attachment (right) efficiency was determined for all strains as described for Fig. 5D,E. For all data n = 3 biological replicates × 3 experimental replicates and p-value was estimated by two tailed Student’s t-test. Error bars show standard deviations (SD). * indicates p-value < 0.05; ** indicates p-value < 0.01; *** indicates p-value < 0.001.

Figure 8

TMT quantification of change in phosphosite abundance plotted against significance of change for 18115 phosphosites in Par versus Δppm5c parasites. Phosphosites significantly more abundant upon PPM5C depletion (P-value < 0.05, log2 fold change Par/Δppm5c < −1, n = 3) and complemented upon reintroduction of PPM5C are highlighted in red. See Supplemental Dataset 2 for full data set.