The exocyst is required for trypanosome invasion and the repair of mechanical plasma membrane wounds

The protozoan parasite Trypanosoma cruzi utilizes the plasma membrane repair machinery for host cell invasion (Fernandes et al., 2011). This machinery includes proteins involved in Ca²⁺-dependent lysosomal exocytosis (Tardieux et al., 1992) and a rapid form of endocytosis triggered by the release of lysosomal acid sphingomyelinase (ASM) (Fernandes et al., 2011; Idone et al., 2008; Tam et al., 2010). Because T. cruzi enters cells in a polarized fashion at the cell periphery, through a process that requires exocytosis (Reddy et al., 2001), we investigated the involvement of the exocyst, a conserved octameric protein complex composed of Sec3, Sec5, Sec6, Sec8, Sec10, Sec15, Exo70 and Exo84 (also known as EXOC1–EXOC8) (Guo et al., 2000; Heider and Munson, 2012; TerBush et al., 1996; TerBush and Novick, 1995), which spatially targets vesicles to the plasma membrane (Boyd et al., 2004; He and Guo, 2009; Moskalenko et al., 2003; Liu and Guo, 2012.).

Exo70 and Sec8 were observed to be associated with sites of contact of extracellular T. cruzi with the host plasma membrane, as identified by the staining of exposed parasite portions with antibodies specific to T. cruzi. Recruitment of Exo70 and Sec8 was markedly stronger around internalized parasite regions (Fig. 1A,B).

The exocyst complex regulates the exocytosis of post-Golgi vesicles (Guo et al., 1999; Guo et al., 2000; He et al., 2007; Hsu et al., 2004; Liu et al., 2007), so we investigated whether post-Golgi vesicles contribute to the formation of the T. cruzi parasitophorous vacuole. HeLa cells were transduced with a construct encoding temperature-sensitive VSVG-ts045–YFP before infection. At the restrictive temperature (39°C) VSVG-ts045–YFP is misfolded and retained in the endoplasmic reticulum. Shifting cells to 37°C allows correct folding and trafficking of the protein through the Golgi to the plasma membrane. Cells were infected with T. cruzi at increasing periods after shifting to 37°C, fixed and imaged by confocal microscopy. Recently internalized parasites (inaccessible to anti-T. cruzi antibodies) were not associated with VSVG-ts045–YFP at any time-point (Fig. 1C), indicating that Golgi-derived membrane traffic is not directed to nascent parasitophorous vacuoles. This new finding complements earlier studies suggesting that intracellular membrane that is recruited to form T. cruzi-containing vacuoles is largely derived from the endosomal or lysosomal compartment (Fernandes et al., 2011; Tardieux et al., 1992; Woolsey et al., 2003).

RNA interference (RNAi)-mediated depletion of Exo70 or Sec8 caused a significant decrease in the number of intracellular parasites when compared with that of cells treated with control siRNA (Fig. 1D,E). Collectively, these results suggest that recruitment of the exocyst complex to nascent parasitophorous vacuoles might regulate localized membrane addition at specific peripheral sites associated with T. cruzi entry.

The unique invasion process utilized by T. cruzi requires exocytic and endocytic events driven by cytosolic Ca²⁺ transients, which are triggered when trypomastigotes initiate contact with host cells (Woolsey et al., 2003; Fernandes et al., 2011). Ca²⁺ signaling promotes recruitment and fusion of host lysosomes at the invasion site, a process involved in T. cruzi invasion of host cells (Tardieux et al., 1992). Thus, our results suggested that the exocyst might promote exocytosis of lysosomes at the invasion site, facilitating parasite invasion. To determine whether the exocyst was involved in generalized lysosomal exocytosis triggered by intracellular Ca²⁺ transients, we measured exocytosis of the lysosomal enzyme β-hexosaminidase in cells depleted of Exo70 and exposed to Ca²⁺ influx by permeabilization with the pore-forming toxin SLO (Fig. 2A), scraping from the substrate (Fig. 2B) or treatment with the Ca²⁺ ionophore ionomycin (not shown). Under all three conditions, Exo70-deficient cells showed comparable or higher levels of β-hexosaminidase secretion in response to Ca²⁺ influx, as compared with controls. This result suggests that the inhibition of T. cruzi invasion in cells depleted in Exo70 and Sec8 is not a consequence of an intrinsic defect in Ca²⁺-triggered lysosomal exocytosis. However, these results do not rule out the possibility that the exocyst complex is required for the highly localized lysosome recruitment and fusion events observed at T. cruzi invasion sites (Tardieux et al., 1992). The increase in generalized Ca²⁺-dependent lysosomal exocytosis in Exo70-depleted cells might be related to changes in the cortical actin cytoskeleton, consistent with the demonstration that Exo70 acts as a kinetic activator of Arp2/3, promoting actin filament nucleation and branching (Liu et al., 2012).

Given that the exocyst plays a role in spatially targeting vesicles to the plasma membrane (Heider and Munson, 2012), we also investigated whether the depletion of exocyst components had a differential effect on the repair of polarized (mechanical) and non-polarized (pore-forming toxin SLO) wounds. After SLO exposure, cells were incubated under conditions that were permissive (Ca²⁺) or non-permissive (no Ca²⁺) for plasma membrane repair, and exposed to the membrane-impermeant dye propidium iodide. Cells treated with control or Exo70-specific siRNA were equally capable of resealing their plasma membrane after SLO permeabilization in the presence of Ca²⁺, as indicated by the block in uptake (Fig. 2C, right panels). As expected, without Ca²⁺, permeabilized cells did not reseal, allowing propidium iodide influx (Fig. 2C, left panels). Flow cytometric quantification of propidium-iodide-positive cells confirmed that depletion of Exo70 (Fig. 2D) or Sec8 (Fig. 2E) did not affect the ability of the cells to repair their plasma membrane after permeabilization with the toxin SLO, which binds to cholesterol and forms stable transmembrane pores with an external diameter of ∼30 nm (Tweten, 2005).

In contrast, Exo70-depleted cells that were mechanically wounded by scraping from the substrate were not able to block propidium iodide influx in the presence of Ca²⁺, reflecting a defect in plasma membrane repair (Fig. 3A). Flow cytometric quantification of propidium-iodide-positive cells after scraping (Fig. 3B,C) confirmed this result, showing that silencing of Exo70 (Fig. 3B, lower right panel) or Sec8 (Fig. 3C, lower right panel) resulted in a strong defect in Ca²⁺-dependent plasma membrane repair. To examine the effect of mechanical wounding on the distribution of Exo70, we injured cells by contact with glass beads in the presence or absence of extracellular Ca²⁺, and performed immunofluorescence with antibodies against endogenous Exo70. Wounded cells were identified by intracellular staining with a membrane-impermeant fluorescent dextran added to the medium during wounding. In non-wounded cells, Exo70 showed a typical punctate intracellular pattern, with minimal association with the plasma membrane, and the presence or absence of Ca²⁺ had no impact on this steady-state distribution (Fig. 4A–D). However, when the cells were wounded with glass beads for 30 s or 3 min, Exo70 was redistributed to the plasma membrane (Fig. 4E–M).

Collectively, our results show that the exocyst tethering complex is recruited to sites of mechanical wounding and of T. cruzi entry, and that this is required for plasma membrane repair and for cell invasion by the parasites. This suggests that the exocyst, which has been previously linked to constitutive secretion, can also regulate Ca²⁺-dependent membrane trafficking events necessary for plasma membrane remodeling. Interestingly, previous studies implicated the exocyst in neurite outgrowth (Murthy et al., 2003), a polarized process that involves the lysosomal SNARE TI-VAMP (also known as VAMP7) (Martinez-Arca et al., 2000) and the lysosomal Ca²⁺ sensor synaptotagmin 7 (Arantes and Andrews, 2006). Thus, the role of the exocyst in T. cruzi invasion and in resealing of mechanical wounds might be attributed to a need for highly localized, spatially regulated lysosomal exocytosis, similar to the process of neurite outgrowth. We cannot rule out, however, that non-lysosomal sources of intracellular membrane contribute to these processes in an exocyst-dependent manner. Regardless of the origin of the exocyst-regulated secretory vesicles, our findings reveal functional differences between the repair of large localized mechanical wounds and smaller homogenously distributed pores generated by the cholesterol-binding toxin SLO. Thus, the resealing of mechanical wounds and T. cruzi invasion might both require polarized exocyst-mediated membrane delivery because they involve events that are restricted to specific plasma membrane sites.

HeLa CCL-2.1 and normal rat kidney (NRK) cells were grown in Dulbecco's modified Eagle's medium (DMEM) (Gibco) supplemented with 10% FBS at 37°C under 5% CO₂. Trypomastigotes from the Trypanosoma cruzi Y strain were obtained from the supernatant of infected monolayers of LLC-MK₂ cells as described previously (Andrews et al., 1987).

1.8×10⁵ cells were plated in 35-mm wells on glass coverslips 24 h before experiments and incubated with 1×10⁸ T. cruzi trypomastigotes in 2 ml of DMEM (containing 2% FBS) per well for various periods of time at 37°C, washed five times with PBS to remove extracellular parasites, fixed with 4% paraformaldehyde (PFA) in PBS for 15 min and processed for inside/outside immunofluorescence (Tardieux et al., 1992). Intracellular parasites were quantified in ∼300 host cells per coverslip in triplicate, by using a Nikon E200 microscope with a 100× NA 1.3 oil objective. For the staining of exocyst components, cells were permeabilized with 0.2% Triton X-100 for 5 min and incubated with mouse monoclonal antibodies (mAb) against Exo70 (Zuo et al., 2006) or Sec8 (Liu et al., 2009) (provided by Dr Wei Guo, University of Pennsylvania, PA) for 1 h, followed by incubation with anti-mouse-IgG secondary antibodies conjugated to Alexa Fluor 488 (Invitrogen). Images were acquired on a laser-scanning confocal Leica SPX5 with a 63× NA 1.4 oil objective. DNA in all samples was stained with 10 µM DAPI (Sigma).

Cell monolayers (at 60% confluence) were washed three times at 4°C with Ca²⁺-free DMEM (containing Mg²⁺ and 10 mM EGTA) and treated with histidine-tagged SLO (R. Tweeten, University of Oklahoma, OK) in Ca²⁺-free DMEM for 5 min at 4°C. Pore formation was triggered by replacing the medium with 37°C DMEM with or without 1.8 mM Ca²⁺. After 3 min, cells were stained with 50 µg/ml propidium iodide (Sigma) for 1 min. For flow cytometry assays, 1×10⁶ trypsinized cells were incubated with SLO for 5 min at 4°C in 250 µl of Ca²⁺-free DMEM, followed by incubation at 37°C in DMEM with or without Ca²⁺ and propidium iodide for 4 min. For scrape assays, cells were removed from the dish at 37°C with a rubber policeman (BD Biosciences). Propidium iodide was added during scraping or after 4 min at 37°C in the presence or absence of Ca²⁺. Flow cytometry data from ∼10,000 cells (FACSCanto, BD Biosciences) were analyzed using FlowJo 6.3 software (Tree Star, Inc.). For scratch assays, monolayers attached to coverslips were scratched with a needle in DMEM with or without Ca²⁺, incubated for 3 min at 37°C, stained for 1 min with propidium iodide and DAPI, and imaged. β-hexosaminidase secretion assays were performed as described previously (Rodríguez et al., 1997).

NRK cells were cultured on glass coverslips to confluency incubated in 500 µl of DMEM with or without Ca²⁺ at 37°C before the medium was exchanged for 500 µl of DMEM with or without Ca²⁺ containing 5 mg/ml Texas-Red–dextran (molecular mass of 10,000 Da, lysine fixable; Invitrogen) and cells were mechanically wounded using 500-µm acid-washed glass beads (Sigma) sprinkled (0.05 g per coverslip) from a 1-ml conical tube held 1 cm above the well. Cells were rocked three times and incubated in DMEM with or without Ca²⁺ or dextran for various time-points, washed in ice-cold PBS without Ca²⁺ and fixed in 4% PFA. Following immunostaining for Exo70, cells were imaged using a Deltavision Elite deconvolution microscope equipped with an NA 1.42 60× objective.

HeLa cells (1.4×10⁵) in 35-mm wells were transfected with Lipofectamine and 160 pmol of stealth siRNA duplexes according to the manufacturer's instructions (Invitrogen). The duplexes used were control (medium GC content; Invitrogen), 5′-AUAGUGAGAGGUAUCAAGG-3′ (targeting human Sec8), 5′-UCGCAGAGAAGAAUCUACCUGUGUU-3′ (targeting human Exo70), 5′-GGCUAAAGGUGACUGACUA-3′ (targeting rat Exo70) or 5′-AUUCCUUGAUGCCACUCAA-3′ (targeting rat Sec8). Cells were used for T. cruzi invasion or plasma membrane repair assays at 72 h post-transfection. Western blotting for Exo70 or Sec8 was performed on cell lysates that were separated by SDS-PAGE, transferred to nitrocellulose and incubated with mouse anti-Exo70 or mouse anti-Sec8 antibodies.

HeLa cells transduced with adenovirus encoding VSVG-ts045–YFP as described previously (Toomre et al., 2000) were placed at 39°C overnight, shifted to 37°C for 15, 30, 45 and 60 min, exposed to trypomastigotes for 10 min, fixed and processed for the quantification of intracellular parasites.

We thank Dr D. Toomre (Yale University) for the VSVG construct, Dr W. Guo (University of Pennsylvania) for antibodies and A. Beaven and K. Class (University of Maryland) for assistance with confocal microscopy and flow cytometry, respectively.