Host microtubule plus-end binding protein CLASP1 influences sequential steps in the Trypanosoma cruzi infection process

Authors


For correspondence. E-mail bburleig@hsph.harvard.edu; Tel. (+1) 617 432 2495; Fax (+1) 617 738 4914.

Summary

Mammalian cell invasion by the protozoan parasite Trypanosoma cruzi involves host cell microtubule dynamics. Microtubules support kinesin-dependent anterograde trafficking of host lysosomes to the cell periphery where targeted lysosome exocytosis elicits remodelling of the plasma membrane and parasite invasion. Here, a novel role for microtubule plus-end tracking proteins (+TIPs) in the co-ordination of T. cruzi trypomastigote internalization and post-entry events is reported. Acute silencing of CLASP1, a +TIP that participates in microtubule stabilization at the cell periphery, impairs trypomastigote internalization without diminishing the capacity for calcium-regulated lysosome exocytosis. Subsequent fusion of the T. cruzi vacuole with host lysosomes and its juxtanuclear positioning are also delayed in CLASP1-depleted cells. These post-entry phenotypes correlate with a generalized impairment of minus-end directed transport of lysosomes in CLASP1 knock-down cells and mimic the effects ofdynactin disruption. Consistent with GSK3β acting as a negative regulator of CLASP function, inhibition of GSK3β activity enhances T. cruzi entry in a CLASP1-dependent manner and expression of constitutively active GSK3β dampens infection. This study provides novel molecular insights into the T. cruzi infection process, emphasizing functional links between parasite-elicited signalling, host microtubule plus-end tracking proteins and dynein-based retrograde transport. Highlighted in this work is a previously unrecognized role for CLASPs in dynamic lysosome positioning, an important aspect of the nutrient sensing response in mammalian cells.

Introduction

Microtubules form polarized dynamic networks in cells that are responsive to extracellular signalling inputs. At the plus-ends, microtubule-associated proteins, referred to as plus-end tracking proteins (+TIPs), assemble as heterogeneous complexes, in a dynamic fashion, to influence tubulin polymer assembly and/or disassembly and to regulate the interactions of microtubules with the cell cortex and other cellular structures (recently reviewed in Galjart, 2010; Jiang and Akhmanova, 2011; Kumar and Wittmann, 2012). The core components of microtubule plus-ends are members of an evolutionarily conserved ‘end-binding’ (EB) protein family. EB1 and EB3 accumulate uniformly at the growing tip of microtubules (Tirnauer and Bierer, 2000) and mediate the binding of other plus-end tracking proteins such as CLIP-170 (Bieling et al., 2008; Wittmann, 2008), cytoplasmic linker-associated proteins (CLASPs) (Akhmanova et al., 2001; Komarova et al., 2005; Mimori-Kiyosue et al., 2005) and the large subunit of dynactin (p150Glued) (Berrueta et al., 1999). Mammalian CLASPs bind to microtubule plus-ends and play a key role in stabilizing microtubules, impacting mitotic spindle formation (Mimori-Kiyosue et al., 2006) and dynamic processes at the cell periphery (Akhmanova et al., 2001; Mimori-Kiyosue et al., 2005). In epithelial cells, CLASPs undergo asymmetric distribution to accumulate at the leading edge where their interactions with components of the cell cortex regulate cytoskeletal dynamics (Wittmann and Waterman-Storer, 2005; Watanabe et al., 2009). EB1 and CLIP-170 localize exclusively to microtubule plus ends, but the interaction of CLASPs with microtubules is spatially regulated. In the main body of migrating cells CLASP binding to microtubules is plus end-restricted whereas CLASPs bind along the length of microtubules in the lamella (Wittmann and Waterman-Storer, 2005). CLASPs regulate peripheral microtubule and actin dynamics via their interactions with actin (Tsvetkov et al., 2007) and regulatory proteins, such as IQGAP (Watanabe et al., 2009) and cell cortex-associated proteins LL5β and ELKS (Lansbergen et al., 2006). Both the microtubule and IQGAP-binding activities of CLASPs are negatively regulated by phosphorylation at multiple sites by glycogen synthase kinase 3β (GSK3β) (Akhmanova et al., 2001; Kumar et al., 2009; Watanabe et al., 2009). Phosphorylation of GSK3β on Ser9 by the serine/threonine kinase, Akt inhibits its kinase activity (Pandey et al., 1998). As such, CLASP-mediated microtubule dynamics at the plus-ends are likely to be responsive to signals that activate PI-3K/Akt signalling in mammalian cells (Akhmanova et al., 2001; Lansbergen et al., 2006).

The microtubule network in mammalian cells is exploited by a variety of intracellular pathogens to facilitate their uptake and for formation, stabilization and maintenance of their intracellular vacuoles (Oelschlaeger et al., 1993; Clausen et al., 1997; Schwan et al., 2009; Aiastui et al., 2010). The protozoan parasite, Trypanosoma cruzi invades a wide variety of non-professional phagocytic cell types where cell invasive trypomastigotes subvert a universal calcium-dependent plasma membrane repair process for entry (Reddy et al., 2001). Ca2+-dependent exocytosis of host lysosomes in the vicinity of the parasite attachment site primes the host plasma membrane for invagination and parasite entry (Tardieux et al., 1992; Rodriguez et al., 1996; 1997; Reddy et al., 2001; Fernandes et al., 2011). Lysosome fusion with the parasite vacuole can coincide with its formation during the T. cruzi internalization process (Tardieux et al., 1992), but more typically, lysosomes fuse with the parasite vacuole a post-entry step in the infection process (Woolsey et al., 2003). Irrespective of differences in timing, lysosome fusion with the T. cruzi vacuole is essential for retention of internalized parasites and thus, for successful establishment of intracellular infection (Andrade and Andrews, 2004; Woolsey and Burleigh, 2004). Host microtubule dynamics are important for facilitating T. cruzi entry into non-professional phagocytes where they have been implicated in the targeting of lysosomes to the plasma membrane during T. cruzi entry (Rodriguez et al., 1996; Tyler et al., 2005). Live cell imaging reveals transient enrichment of GFP-α-tubulin at the parasite attachment site (Tyler et al., 2005) and the directed movement of peripheral lysosomes towards the invading parasite (Rodriguez et al., 1996). Disruption of anterograde lysosome transport in host cells with microinjected antibodies to kinesin heavy chain impedes cell entry by T. cruzi, as does chemical inhibition of host microtubule dynamics (Rodriguez et al., 1996). Together, these observations support a model in which T. cruzi trypomastigotes exploit kinesin-based motility to transport host lysosomes to the cell periphery, where they undergo regulated exocytosis at the parasite attachment site to facilitate entry and vacuole biogenesis. Cytosolic free calcium transients triggered in host cells by invasive T. cruzi trypomastigotes (Tardieux et al., 1994; Caler et al., 2000) are associated with transient depolymerization of cortical actin microfilaments (Rodriguez et al., 1995) and lysosome exocytosis (Rodriguez et al., 1997; Fernandes et al., 2011). T. cruzi also stimulates the localized production of PI(3,4,5)P3 at the host cell plasma membrane in a class I PI-3K-dependent manner prior to entry, where chemical and genetic targeting of host PI-3K inhibited parasite uptake and lysosome fusion with the nascent parasite vacuole (Woolsey et al., 2003). Despite this basic knowledge, the mechanisms by which T. cruzi-triggered signalling pathways integrate with host membrane and cytoskeletal dynamics to promote parasite entry remain largely unknown. Given the highly responsive nature of the cortical cytoskeleton to extracellular signals (Saarikangas et al., 2010) and the key role played by microtubule plus-end binding proteins in cellular responses to these cues (Akhmanova et al., 2001; Zhou et al., 2004; Lansbergen et al., 2006), we sought to determine whether microtubule plus-end binding proteins participate in the T. cruzi infection process.

Results

T. cruzi trypomastigote invasion is oriented with respect to host microtubules

Previous studies have indicated that host microtubule dynamics contribute to the entry of non-professional phagocytes by T. cruzi tryopmastigotes (Rodriguez et al., 1996; Tyler et al., 2005). It has also been appreciated that T. cruzi trypomastigotes exhibit a strong preference for invasion at cell edges (Schenkman et al., 1988). Using EB3-GFP to decorate microtubule plus-ends in polarized human foreskin fibroblasts (HFF), we observe that T. cruzi trypomastigotes display uniaxial entry at the elongated ends of fibroblasts, where they enter in a ‘head-on’ orientation to the direction of microtubule elongation (Fig. 1A–C). Examination of 200 random cells with invading or recently internalized parasites reveals that > 80% T. cruzi trypomastigotes are oriented with host microtubules in a parallel fashion (Fig. 1D). Following invasion of host cells, T. cruzi transiently reside in tight-fitting vacuoles that accumulate host lysosomal markers within the first 60–90 min (Woolsey et al., 2003). Parasite-containing vacuoles remain LAMP-positive for at least 8 h (Chessler et al., 2008) before their disruption (Andrews et al., 1990; Ley et al., 1990) and asynchronous release of parasites into the host cytosol (Chessler et al., 2008). We noted here that internalized T. cruzi parasites move inward towards the host cell nucleus (Fig. 1E) where approximately half of the internalized parasites exhibit juxtanuclear localization 90 min post-entry (Fig. 1E). Although this number rapidly plateaus early in the infection (Fig. 1E), all of the internalized parasites exhibit juxtanuclear positioning by 24 h (data not shown). The biphasic nature of this process is curious and may reflect heterogeneity in the molecular composition of the parasite vacuole and its ability to interact with host machinery required for trafficking. The significance of juxtanuclear positioning of the T. cruzi vacuole is currently not known; however, the minus-end directed movement of internalized parasites suggests a role for microtubules and dynein–dynactin motor complexes.

Figure 1.

T. cruzi trypomastigotes enter cells in a parallel orientation to host microtubules.

A. Representative image of a recently internalized T. cruzi trypomastigote in HFF with the microtubules with EB3-GFP (green). Double-headed arrow indicates the orientation of the main microtubules in elongated HFF. Single arrow indicates direction of parasite entry into cell and highlights the intracellular parasite.

B. Higher magnification of recently internalized trypomastigote shown in (A); outline of the parasite as it displaces microtubules (green) (upper) is seen in the DIC image (lower).

C. Image of HFF expressing EB3-GFP (green) enriched at the growing microtubule plus-ends (comet like structures; white arrowheads) with invading T. cruzi trypomastigote [DAPI-stained (blue) kinetoplast (k) and nucleus (n) indicated with arrowhead and parasite length bracketed by red bars]. Long arrow indicates the orientation of microtubule elongation in HFF.

D. The orientation of invading or recently internalized parasites relative to host cell microtubules was scored in 200 infected cells.

E. The relative number of intracellular T. cruzi exhibiting juxtanuclear localization at indicated time points (graph). HFF were pulsed with trypomastigotes (50 parasites per host cell) for 15 min, washed to remove extracellular parasites and incubated for the indicated times prior to aldehyde fixation and immunostaining with anti-T. cruzi antibody (green) to detect extracellular parasites and counterstained with anti-LAMP1 (red) and DAPI (blue). Graphical data are expressed as the mean ± SD for triplicate samples representative of three independent experiments. Representative images are shown to the right.

CLASP1 depletion impairs T. cruzi internalization and delays lysosome fusion with the vacuole

Given the central role for microtubule plus-end tracking proteins (+TIPs) in regulating microtubule dynamics and their responsiveness to cellular signalling cues known to regulate T. cruzi entry, we sought to determine whether host microtubule +TIPs function in the establishment of intracellular T. cruzi infection. We focused on CLASPs as peripheral +TIPs that bridge interactions between microtubule plus-ends and the cell cortex (Akhmanova et al., 2001). Enrichment of GFP-CLASP1/2 at the site of T. cruzi trypomastigote entry and surrounding recently internalized parasites in time-lapse imaging of live cells (Fig. 2; Fig. S1) is consistent with previous observations of transient localization of GFP-α-tubulin at the parasite attachment site (Tyler et al., 2005). The relative enrichment of CLASP-GFP with invading or recently internalized T. cruzi trypomastigotes as compared with cytosolic GFP was quantified (Fig. 2E; Fig. S1).

Figure 2.

CLASP is transiently enriched at the T. cruzi invasion site.

A and B. Time-lapse fluorescence images of recently internalized T. cruzi trypomastigotes (outlined in white in first frames) in HFF expressing CLASP1-GFP (A) or CLASP2-GFP (B). Image capture was initiated 10 min after initial incubation of parasites with host cells (marked as 0 seconds). Arrows indicate sites of enriched CLASP1-GFP or CLASP2-GFP. Scale bar = 5 μm.

C. Quantification of CLASP1-GFP enrichment with invading or recently internalized trypomastigotes as determined by the Linescan function in the MetaMorph® imaging software and compared with cytosolic GFP expression. Measurements were made for 25–30 cells each of control (GFP), CLASP1-GFP and CLASP2-GFP and reported as the ratios of pixel intensity for GFP signal proximal to parasites/pixel intensity for non-parasite-associated GFP signal along the Linescan measurement (refer to Supporting Fig. S1). Data were analysed using one-way anova; *P < 0.01.

Acute siRNA-mediated silencing of CLASP1 in HeLa cells (Fig. 3A) significantly compromised the capacity for T. cruzi trypomastigote invasion (Fig. 3B) without affecting Ca2+-triggered lysosome exocytosis, as measured by the release of β-hexosaminidase in the presence of Ca2+ in streptolysin-O-permeabilized cells (Fig. S2). This reduced capacity for parasite invasion in CLASP1-depleted cells was replicated with all four of the individual siRNA duplexes represented in the original siRNA CLASP1-targeting pool and with an independent pool with different targeting sequences (Fig. S3), providing confidence that the phenotypic effects of CLASP1 depletion are specific (i.e. not owing to an ‘off-target’ effect). In addition to the impact of CLASP1-depletion on T. cruzi trypomastigote entry, the subsequent delivery of the lysosomal marker, LAMP-1, to the parasite vacuole was significantly delayed (Fig. 3C). Despite observations that CLASP1-GFP and CLASP2-GFP are enriched at the site of parasite entry and around recently internalized parasites (Fig. 2), acute silencing of endogenous CLASP2 (Fig. S4A), which shares a high degree of homology with CLASP1 with reported overlapping/redundant functions (Akhmanova et al., 2001; Mimori-Kiyosue et al., 2005), failed to impact T. cruzi invasion and LAMP1 delivery to the parasite-containing vacuole (Fig. S4B and C). A notable difference between CLASP1- and CLASP2-silenced cells is the reduction of actin stress fibres that is consistently associated with CLASP1 knock-down (Fig. S4D and data not shown). As CLASP1 transcript levels are approximately threefold higher than CLASP2 transcripts in siControl-transfected HeLa cells (Fig. S4A) it is possible that the phenotypic effects associated with CLASP1 silencing are only observed when a critical threshold of CLASP is reached in the cell and that targeting CLASP2 fails to lower the total CLASP to this threshold. The observation that cell infection by Toxoplasma gondii, a protozoan parasite with a completely different mode of entry, was not impaired in CLASP1-depleted cells (Fig. S5) suggests that cellular phenotypes associated with CLASP1 knock-down are related to the specific mode of host cell entry exploited by T. cruzi trypomastigotes (Caradonna and Burleigh, 2011).

Figure 3.

CLASP1 depletion inhibits T. cruzi invasion and subsequent fusion with host cell lysosomes. siRNA-mediated silencing of endogenous CLASP1 (A), DCTN1 (D) and EB1 (G) 48 h post transfection is demonstrated by Western blots to detect protein with specific antibodies. Relative T. cruzi invasion after 15 min (B, F and H) or LAMP-1 association with the parasite vacuole (C, G and I) in HeLa cells following siRNA-mediated silencing of CLASP1 (A–C), DCTN1 (D–F) or EB1 (G–I) and compared with cells transfected with non-targeting siRNA (siControl). Data are represented as mean ± SD, n = 3 analysed by Student's t-test. *P < 0.05.

To relate the CLASP1 phenotype to microtubule dynamics at the plus-ends we examined the impact of silencing the end-binding protein, EB1 on T. cruzi entry and kinetics of LAMP1 delivery to the parasite vacuole. EB1 is considered a ‘core’ component of the +TIPs, since virtually all other +TIPs, including CLASP1, bind EB1 directly and require EB1 for plus-end localization, whereas EB1 itself can track microtubule ends autonomously. Unlike CLASP1, acute silencing of EB1 (Fig. 3D) failed to affect T. cruzi entry (Fig. 3E) and did not significantly impact LAMP1 accumulation in the parasite vacuole (Fig. 3F). Similar findings were observed when EB2 and EB3 [homologues of EB1 (Lansbergen and Akhmanova, 2006)] were targeted in HeLa cells, singly or in combination with EB1 (data not shown). We next tested the role for dynactin, a large multi-protein complex that interacts with microtubules and dynein motors, which has been implicated in microtubule-dependent vesicular trafficking (Burkhardt et al., 1997; Valetti et al., 1999; Habermann et al., 2001; Vaughan, 2005). A core component of dynactin, p150Glued (DCTN1), colocalizes with EB1 at microtubule plus-ends and facilitates loading of vesicular cargo onto dynein (Vaughan et al., 2002). While acute silencing of DCTN1 (Fig. 3G) failed to inhibit T. cruzi entry (Fig. 3H) a marked delay in LAMP1 delivery to the parasite vacuole was observed (Fig. 3I) similar to the effect of CLASP1 depletion (Fig. 3C).

+TIPs modulate the juxtanuclear positioning of the T. cruzi vacuole

Following host cell entry, internalized T. cruzi migrate towards the host nucleus where they undergo a developmental switch and eventually lyse the vacuole to become cytoplasmically localized (Andrews et al., 1990). To assess the impact of +TIP depletion on the inward movement of the T. cruzi vacuole, HFF transfected with control non-targeting siRNA or siRNAs targeting CLASP1 or EB1, were pulsed with T. cruzi trypomastigotes for 15 min and the juxtanuclear localization of parasite-containing vacuoles was evaluated at 1.5 h post infection, a time point at which approximately half of the internalized parasites have migrated to the nucleus (Fig. 1E). Fewer internalized parasites achieved a juxtanuclear position by 1.5 h in EB1- or CLASP1-silenced cells (Fig. 4A–C) or in cells overexpressing p50 dynamitin, a subunit of the dynactin complex which disrupts dynein-dependent cargo transport (Burkhardt et al., 1997) (Fig. 4D and E). In agreement with other phenotypic differences observed between CLASP1- and CLASP2-depleted cells, acute silencing of CLASP2 fails to delay juxtanclear positioning of internalized T. cruzi (Fig. S4E). Together, these data highlight the critical influence of CLASP1 expression on T. cruzi trypomastigote entry as well as the post-entry steps of lysosome–vacuole fusion and juxtanuclear positioning of the parasite vacuole and suggest a potential coupling of entry and post-entry events in the parasite infection process. Unlike CLASP1, the core microtubule end-binding protein EB1 and dynactin were dispensable for T. cruzi internalization. However, dynactin influenced the post-entry steps of lysosome–vacuole fusion and the inward migration of the parasite-containing vacuole as did EB1.

Figure 4.

A. Western blots to detect EB1 and CLASP1 protein expression at 48 h post siRNA-mediated silencing in HFF cells as compared to cells transfected with non-targeting control siRNA (siControl).

B. Graphical representation of relative numbers of intracellular parasites that exhibit juxtanuclear localization in HFF 1.5 h post-internalization following siRNA-mediated silencing of EB1 (siEB1) or CLASP1 (siCLASP1) or in cells transfected with non-targeting control siRNA (siControl). Transfected cells were pulsed with T. cruzi trypomastigotes for 15 min followed by a ‘chase’ in medium for 1.5 h.

C. Representative images of graph B showing where transfected cells were identified with a fluorescent siRNA tracker (red) and intracellular parasites identified by DAPI staining (blue).

D. Dynactin p50-GFP expression delays juxtanuclear migration of T. cruzi (DAPI: blue). Data are represented as graph D as means ± SD, n = 3; and analysed by Student's t-test. *P < 0.05; and representative images in (E).

CLASPs influence dynamic positioning of lysosomes in mammalian cells

Our observation that depletion of CLASP1 or dynactin (DCTN1) in mammalian cells exerts similar effects on the post-entry processes of lysosome marker delivery to the T. cruzi vacuole and its retrograde transport towards the host cell nucleus, leads us to investigate whether CLASP1 expression influences minus-end directed lysosome transport, a dynein-dependent process (Jordens et al., 2001). To test this, we exploited the ability to induce rapid redistribution of lysosomes by altering cytosolic pH (Heuser, 1989). At neutral pH, lysosomes in control and CLASP1-depleted cells display similar distributions (Fig. 5; pH 7.2). However, the rapid perinuclear clustering of lysosomes triggered by a shift to alkaline pH conditions is significantly impaired in CLASP1 knock-down cells (Fig. 5; pH 7.6). Similar results were obtained when starvation was used as the stimulus to promote perinuclear localization of lysosomes (data not shown) (Korolchuk et al., 2011). Here we show that lysosome redistribution under alkaline conditions is also compromised in CLASP2-depleted cells as compared with control cells (Fig. S6) however the impairment appears to be less pronounced than in CLASP1-silenced cells. These novel findings indicate that in addition to its role in microtubule plus-end stabilization, CLASPs plays an important role in microtubule-dependent lysosome positioning in cells similar to that observed for dynein or p150Glued (Caviston et al., 2011; Tan et al., 2011).

Figure 5.

CLASP1 is required for minus-end directed lysosome trafficking. Representative images of HeLa cells transfected with non-targeting siRNA (siControl) or CLASP1-specific siRNA (siCLASP1) following a 15 min exposure to Ringer's-BSA pH 7.2 or pH 7.6 to induce perinuclear localization of lysosomes. Aldehyde-fixed and permeabilized cells were stained with anti-LAMP1 (red) and counterstained with DAPI (blue). Quantification of perinuclear lysosome clustering for 200 cells (right panel). Data represented as mean ± SD, n = 3; and analysed by Student's t-test. *P < 0.01.

Inactivation of GSK3β increases T. cruzi entry

CLASP recruitment to microtubule plus-ends and its interaction with EB1 and IQGAP are negatively regulated by GSK3β-dependent phosphorylation at multiple sites (Kumar et al., 2009; Watanabe et al., 2009). While the majority of the work characterizing these interactions have been focused on CLASP2, homology between CLASP1 and CLASP2 and the conservation of putative GSK3β phosphorylation sites, suggests that CLASP1 is also negatively regulated by GSK3β. The activity of GSK3β is inhibited when phosphorylated on Ser9 by the serine/threonine kinase Akt (Jope and Johnson, 2004) which becomes activated in a PIP3-dependent manner downstream of class I PI-3Ks (Coffer et al., 1998). The elevated levels of phospho-AktSer473 and phospho-GSK3βSer9 triggered in response to T. cruzi (Fig. 6A) are consistent with the ability of these parasites to exploit host PI-3K-dependent signalling for invasion (Woolsey et al., 2003). Consistent with a role for CLASP1 in T. cruzi entry, chemical inhibition of GSK3β, predicted to increase CLASP1 functionality, enhanced trypomastigote invasion (Fig. 6B) in a CLASP1-dependent manner (Fig. 6C). Conversely, expression of constitutively active GSK3β-S9A, which is known to inhibit CLASP binding of microtubule plus-ends (Akhmanova et al., 2001; Wittmann and Waterman-Storer, 2005), hinders T. cruzi invasion (Fig. 6D). Thus, with its clear influence on T. cruzi internalization and functional regulation downstream of the Akt-GSK3β signalling axis, CLASP1 is well positioned to integrate parasite-elicited signals with host microtubule dynamics to influence T. cruzi internalization. Moreover, the extended influence of CLASP1 on post-entry steps in the T. cruzi infection process suggests that cellular processes involved in the establishment of intracellular T. cruzi infection, normally considered as distinct (internalization, vacuole biogenesis/maturation, vacuole trafficking), are likely to exploit overlapping molecular mechanisms.

Figure 6.

Inactivation of host GSK3β enhances T. cruzi entry.

A. Immunoblots demonstrating increased phosphorylation of host cell AktSer473 and GSK3βSer9 in lysates of mock (C) or T. cruzi-infected (T.c.) HeLa or HFF infected for 15 min. Blots were stripped and reprobed with antibodies to total Akt and β-actin to normalize signals.

B. Pre-treatment of HeLa with GSK3β inhibitor (GSKi) for 30 min enhances T. cruzi invasion.

C. GSKi pre-treatment of HeLa enhances internalization of T. cruzi trypomastigotes cells transfected with non-targeting siRNA (siControl) but fails to affect invasion of cells lacking CLASP1 (siCLASP1).

D. T. cruzi invasion is reduced in cells expressing constitutively active GSK3β-S9A. The relative number of internalized T. cruzi was determined in cells infected with parasites for 15 min (B–D).

Data are represented as mean ± SD. n = 3; and analysed by Student's t-test. *P < 0.05.

Discussion

Establishment of intracellular infection by T. cruzi trypomastigotes is critical to the pathogenesis of human Chagas' disease, where the importance of host microtubule dynamics in the process of T. cruzi entry into non-professional phagocytic cells has been documented (Tardieux et al., 1992; Rodriguez et al., 1996; Tyler et al., 2005). Host microtubules associate with invading T. cruzi trypomastigotes and their nascent vacuoles (Tyler et al., 2005). Microtubules function as a conduit for the kinesin-dependent, anterograde transport of lysosomes to the parasite attachment site (Rodriguez et al., 1996) where regulated lysosome exocytosis facilitates T. cruzi internalization (Tardieux et al., 1992; Rodriguez et al., 1996; 1997; Reddy et al., 2001; Fernandes et al., 2011). Results from the present study expand our view of the role of host microtubules in the T. cruzi infection process, revealing the marked influence of microtubule plus-end tracking proteins on trypomastigote internalization, vacuole biogenesis and retrograde transport of the parasite vacuole (Fig. 7). Accompanying these findings is the demonstration of a previously unrecognized role for CLASPs in dynamic lysosome positioning in mammalian cells, which has important implications for nutrient sensing, autophagy and innate immunity (Korolchuk et al., 2011).

Figure 7.

Integrated model of host microtubule function during the establishment of intracellular T. cruzi infection. (1) To initiate infection, invasive T. cruzi trypomastigotes activate signalling pathways in mammalian host cells including intracellular Ca2+-transients and activation of class I PI-3Ks. Elevated Ca2+ results in transient depolymerization of the cortical actin cytoskeleton and promotes lysosome–plasma membrane fusion. Host microtubules transport lysosomes to the cell periphery in a kinesin-dependent manner. (2) Trypomastigotes invade at ceramide-rich regions of the plasma membrane generated by the action of released acid sphingomyelinase from exocytosed lysosomes. Parasite-triggered activation of PI-3K/Akt results in phosphorylation and inactivation of GSK3β, and increased CLASP1 association with microtubule plus-ends and cortical actin. (3) CLASP1 participates in connecting invading or (4) recently internalized parasite vacuole membranes to host microtubules which function to pull the parasite vacuole membrane towards the nucleus and facilitate fusion with lysosomes.

Acute silencing of CLASP1 results in significant impairment of T. cruzi trypomastigote invasion of mammalian cells without perturbing anterograde movement of lysosomes (data not shown) or the capacity for Ca2+-regulated lysosome exocytosis (Rodriguez et al., 1997). These findings suggest that CLASP1 expression influences events in the T. cruzi invasion process downstream of the kinesin-based delivery of lysosomes to the cell periphery (Rodriguez et al., 1996). Initially puzzling was the observation that silencing of CLASP2, a closely related functional homologue of CLASP1 (Mimori-Kiyosue et al., 2005), fails to replicate the parasite infection phenotypes observed in CLASP1-depleted cells. However, the low level of endogenous CLASP2 mRNA expression in HeLa cells suggests that silencing of CLASP2 does not sufficiently reduce the functional pool of CLASPs in the cell, whereas silencing of the more abundant CLASP1 does. The observation that CLASP2 silencing resulted in a weaker impairment of pH-dependent lysosome distribution in mammalian cells than CLASP1 depletion supports this idea. Alternatively, CLASP1 may exhibit functions that impact the establishment of intracellular T. cruzi infection that are independent of those of CLASP2.

By analysing the consequences of depleting host microtubule plus-end tracking proteins on the T. cruzi infection process, a general hierarchy emerged in which CLASP1 expression impacts both parasite entry and post-entry events, whereas dynactin and EB1 function at steps downstream of parasite internalization (Fig. 7). Disruption of dynactin function, which has a well-established role in the maturation of phagosomes and other pathogen vacuoles (Blocker et al., 1997; Jordens et al., 2001; Harrison et al., 2003) delays lysosome fusion with the T. cruzi vacuole as well as retrograde transport of internalized parasites, similar to that observed for CLASP1 depletion. In contrast, silencing of EB1, which is required for the plus-end localization of dynactin p150Glued via CLIP-170 (Watson and Stephens, 2006) fails to impact the kinetics of lysosome marker delivery to the T. cruzi vacuole. While this observation is consistent with previous reports that dynactin can function independently of microtubule plus-ends (Watson and Stephens, 2006), the involvement of CLASP1, dynactin and EB1 in the juxtanuclear positioning of the parasite vacuole supports a role for microtubule plus-end stabilization in facilitating interaction of the T. cruzi vacuole with dynactin–dynein motor complexes to drive the inward movement of the vacuole. Our finding that CLASP1 depletion impairs the dynein and microtubule-dependent process of perinuclear lysosome aggregation in response to increased cytosolic pH (Heuser, 1989; Caviston et al., 2011), suggests a similar mechanism of organelle/vacuole ‘capture’ and transport (Vaughan et al., 2002). involving the early participation of CLASP1.

The enrichment of CLASP-GFP at the T. cruzi entry site is consistent with its role in the early stages of the parasite internalization process. Given that the role of CLASP1 in T. cruzi invasion can be uncoupled from other microtubule plus-end binding proteins (EB1 and dynactin) points to unique molecular interactions between CLASP1 and other cellular proteins as critical for influencing the parasite internalization process. Dynamic interactions between CLASPs and the actin regulatory protein IQGAP (Watanabe et al., 2009) or cell cortex-associated proteins (LL5beta and ELKS) (Lansbergen et al., 2006) may impact T. cruzi internalization, directly or indirectly, by modulating cortical actin filament dynamics. Although the T. cruzi entry process is frequently referred to as ‘actin-independent’ because cortical actin polymerization is not a requirement for trypomastigote invasion of non-professional phagocytic cells (Schenkman et al., 1991; Tardieux et al., 1992), this description is misleading given that cortical actin depolymerization, which is associated with parasite-elicited Ca2+ transients, is known to facilitate parasite entry (Tardieux et al., 1994; Rodriguez et al., 1995). Moreover, while cytochalasin D-mediated inhibition of actin polymerization initially enhances T. cruzi trypomastigote internalization in many mammalian cell types (Tardieux et al., 1992), recovery from cytochalasin D treatment is essential to promote fusion of early endosomes and lysosomes with the parasite vacuole (Woolsey and Burleigh, 2004). Moreover, with prolonged impairment of actin microfilament dynamics there is significant loss of internalized parasites (Woolsey and Burleigh, 2004). Thus, host actin dynamics are important for T. cruzi trypomastigote invasion and cellular retention (Tardieux et al., 1992; Rodriguez et al., 1995; Woolsey and Burleigh, 2004). As CLASP1 depletion results in loss of actin stress fibres (Fig. S4), we cannot rule out the possibility that the influence of CLASP1 on T. cruzi invasion is related to altered actin microfilament dynamics independent of the host microtubule network.

CLASP binding to microtubules and to IQGAP is negatively regulated by the serine/threonine kinase GSK3β where chemical inhibition of GSK3β enhances CLASP localization to microtubules and stabilizes microtubules at the leading edge of migrating cells (Kumar et al., 2009; Watanabe et al., 2009). Consistent with a role for GSK3β-dependent inhibition of CLASP1, inhibition of GSK3β activity enhances T. cruzi entry in a CLASP1-dependent manner, whereas expression of constitutively active GSK3β dampens infection. Thus, localized activation of class IA PI-3 kinases and PIP3 accumulation at the T. cruzi invasion site (Woolsey et al., 2003) may influence the microtubule and IQGAP-binding activity of CLASPs following Akt-dependent phosphorylation of GSK3β. While we were able to demonstrate that exposure of mammalian cells to T. cruzi trypomastigotes results in increased phosphorylation of GSK3β on Ser9, which would lead to inactivation of the kinase (Jope and Johnson, 2004), we have not been able to detect changes in the phosphorylation state of CLASPs in response to T. cruzi-triggered signalling at a population level despite several attempts (data not shown). Given the transient and asynchronous nature of T. cruzi-elicited signalling in mammalian host cells (Caler et al., 2000; Woolsey et al., 2003), more sensitive methods will be needed to detect changes in the phosphorylation state of CLASP at the single cell level. However, coupled with the fact that IQGAP activities are modulated by Ca2+ calmodulin (Atcheson et al., 2011), and CLASP binding to IQGAP and EB1 is regulated by GSK3β, CLASPs are well positioned both spatially and biochemically to integrate PI-3K and Ca2+-dependent signalling pathways triggered in mammalian host cells early in the T. cruzi entry process. Based on current knowledge, both CLASP1 and CLASP2 should be equally competent in this regard.

Overall, this study provides novel information regarding the role of host microtubule plus-end binding proteins in the establishment of intracellular infection by the protozoan parasite, Trypanosoma cruzi. CLASP1 plays an important role in parasite internalization and in the dyactin-dependent events of vacuole maturation and its minus-end directed movement. These findings suggest that mammalian CLASPs serve as a critical point of integration for T. cruzi-dependent signalling to the actin and microtubule networks in host cells and couple parasite entry with lysosome fusion and juxtanuclear localization of the parasite-containing vacuole. A novel finding of general interest to cell biologists is the role of CLASPs in dynamic positioning of lysosomes in the cell where their minus-end directed movement is impaired in the absence of CLASP1 and to a lesser degree CLASP2. While the role of CLASP1 in the regulation of actin dynamics is not yet understood, recent studies that have identified several new CLASP-interacting proteins in Drosophila are certain to pave the way for more mechanistic studies (Lowery et al., 2010). Exploiting pathogens as perturbagens of the CLASP interactome (Schwan et al., 2009) has the potential to illuminate novel biology and will create opportunities to enhance our understanding of the molecular and cellular basis for establishment of intracellular infection by this important human pathogen.

Experimental procedures

siRNAs, plasmids and adenovirus

Gene-specific siGENOME SMARTpools containing four individual siRNAs targeting human CLASP1, CLASP2, EB1, EB3, DCTN1, a non-targeting control pool siRNA2 (siRNA#2) and individual siRNAs targeting CLASP1 were purchased from Dharmacon (ThermoScientific). siRNAs targeting CLASP1 were: siGENOME SMARTpool M-00683-01-0005 containing four siRNA duplexes, the individual duplexes deconvoluted from this pool: D-006831-01-005, D-006831-02-005, D-006831-04-005, D-006831-17-005 and ON-TARGETplus SMARTpool L-006831-00-0005. EB3-GFP, CLASP1-GFP, CLASP2-GFP plasmids were characterized as described (Akhmanova et al., 2001). p50-GFP plasmid was kindly provided by Dr Vaughan (University of Notre Dame). Constitutively active GSK3β-S9A adenovirus and GFP adenovirus were tested previously (Kumar et al., 2009).

Cell culture, transfection and T. cruzi infection

HeLa and human foreskin fibroblasts (HFF), purchased from the American Type Culture Collection (ATCC), were maintained as subconfluent monolayers in Dulbecco's modified Eagle's medium (DMEM) containing 4.5 g l−1 glucose (Invitrogen) and 10% fetal bovine serum (FBS, Invitrogen). Gene silencing in HeLa cells was achieved using a reverse transfection protocol where 3.5 × 104 HeLa cells were suspended in 115 μl of Opti-MEM (Invitrogen) with 1.7 μl Oligofectamine (Invitrogen) and 25 or 50 nM of a gene-specific or non-targeting siRNA pool. Cells were seeded on 12 mm2 glass coverslips in 24-well plates (Corning), adjusted to 10% FBS in DMEM and incubated at 37°C, 5% CO2 for 48 h prior to infection with T. cruzi. Transient transfection of HFF was achieved by nucleofection using an Amaxa NucleofectorTMII (Lonza) according to the manufacturer's specifications. Briefly, 5 × 105 HFF suspended in transfection buffer (Amaxa human dermal fibroblast nucleofector kit, Lonza) were mixed with 50 nM SMARTpool siRNA or 1 μg of plasmid DNA and immediately seeded onto glass coverslips in six-well plates. To identify transfected cells siRNAs were co-transfected with red fluorescent siRNA tracker (Dharmacon D-001630-02-05). Transfected cells were grown for 48 h prior to T. cruzi infection or cell treatments. To determine knock-down efficiency, Western blots of total-cell lysates were probed with specific antibodies.

Parasite infections

Trypanosoma cruzi trypomastigotes (Tulahuén) isolated from LLcMK2 monolayers as described (Woolsey et al., 2003) were washed in Ringers/BSA or DMEM and incubated with transfected cells for 15 min at 37°C. Cells were washed five times with 2% FBS DMEM and incubated in 2% FBS DMEM for indicated times. To inhibit mammalian GSK3β, HeLa cells were pre-treated with 1 μM of the GSK3β inhibitor BIO (Tocris Bioscience) for 30 min, washed three times with 2% FBS DMEM and infected with T. cruzi for 15 min. Infected cells were fixed in 2% paraformaldehyde/PBS and relative T. cruzi invasion was determined by anti-T. cruzi antibodies and fluorescence dye followed by staining with 1.25 μg ml−1 4′, 6-diamidino-2-phenylindole (DAPI) to visualize host and parasite nuclei as described (Woolsey et al., 2003). To determine the impact of gene depletion on lysosome distribution, transfected HeLa cells were incubated for 15 min in Ringers/BSA solution at pH 7.2 or 7.6 (Heuser, 1989), fixed and lysosomes were visualized following staining with anti-human LAMP1 (The monoclonal antibody H4A3 developed by J.T. August and J.E.K. Hildreth was obtained from the Developmental Studies Hybridoma Bank developed under the auspices of the NICHD and maintained by The University of Iowa, Department of Biology, Iowa City, IA 52242). Fluorescence images were obtained with a Nikon TE-300 inverted epifluorescence microscope equipped with an Orca-100 CCD camera (Hamamatsu) and MetaMorph imaging software (Universal Imaging Corporation). For infections with Toxoplasma gondii, negative control or CLASP1 siRNA-transfected HeLa cells were infected with 4000 β-galactosidase-expressing RH strain Toxoplasma tachyzoites. After 72 h, the medium was removed and 100 μl of Z-buffer [10 mM NaPO4 (pH 7.4), 137 mM NaCl, 2.7 mM KCl, 9 mM MgCl2, 0.125% NP-40, 100 mM β-mercaptoethanol] containing 20 μM CPRG was added to each well. Plates were incubated at 37°C for 15 min and absorbance measured at 570 nm. Numbers of parasites in each well were determined by linear regression analysis from a standard curve prepared in each plate.

Western blot analysis

Cells were lysed in Laemmli sample buffer and total protein separated on 10% SDS-PAGE gels (Bio-Rad) and electrophoretically transferred to PVDF membranes (Millipore). Blots were blocked with 5% non-fat dried milk in TBST (10 mM Tris, 150 mM NaCl, 0.05% Tween-20) for 1 h at room temperature. CLASP1 antibody (Epitomics 1:5000) and EB1 antibody (Millipore 1:2000) were diluted in 5% BSA in TBST (10 mM Tris, 150 mM NaCl, 0.05% Tween-20) and incubated with blots overnight at 4°C. ECL peroxidase-linked secondary antibody (Invitrogen) 1:2000 in blocking buffer was used to detect the signal. To detect phosphorylation Akt (Cell Signaling), endogenous AKT (Cell Signaling) and GSK3β (Cell Signaling), antibodies were diluted 1:1000 in 5% BSA in TBST and incubated with blots overnight at 4°C. Blots were stripped and re-probed with antibodies to α-actin (Sigma; 1:10 000) for normalization of signals. Blots were developed with Supersignal™ substrate (Thermo Scientific).

Lysosome exocytosis

HFF monolayers (60% confluence) were washed at 4°C with Ca2+-free buffer (containing Mg2+ and 10 mM EGTA), followed by two more washes in Ca2+-free buffer. SLO (Sigma-Aldrich) was bound to target cells in Ca2+-free buffer for 5 min at 4°C, and pore formation was triggered by replacing the medium with 37°C buffer containing or not 1.8 mM Ca2+. After 10 min at 37°C, cell supernatants were stained for 15 min with 2.3 μg ml−1 β-hexosaminidase substrate (Sigma-Aldrich) and analysed by spectrofluorimeter as described in Rodriguez et al. (1997).

RNA extraction and quantitative reverse transcription-PCR (RT-PCR)

To determine efficiency of CLASP1 and CLASP2 knock-down, RNA was extracted (Qiagen, 74104) and reverse transcripted to cDNA using iScript (Invitrogen, 170-8890). Human CLASP1 (Hs01076541_m1, Taqman, Invitrogen) and CLASP2 (Hs00380556_m1, Taqman, Invitrogen) were amplified in triplicates by using standard amplification conditions for ABI (ABI PRISM 7900 HTA FAST). GAPDH was used to normalize the Ct values of the target genes.

Acknowledgements

Authors would like to gratefully acknowledge Dr Vaughan for providing the p50-GFP construct and helpful suggestions and D. Ndegwa for excellent technical support. This work was supported by the National Institutes of Health Grant R21AI090366 and Harvard Medical School Milton Fund awarded to B.A.B. K.L.C. was supported by the Bayer Fund for Scholars in Infectious Diseases. I.B. was supported by NIH-R21AI087485 and MBC-114461.

Ancillary