Host Cell Rap1b mediates cAMP-dependent invasion by Trypanosoma cruzi

Trypanosoma cruzi cAMP-mediated invasion has long been described, however, the detailed mechanism of action of the pathway activated by this cyclic nucleotide still remains unknown. We have recently demonstrated a crucial role for Epac in the cAMP-mediated invasion of the host cell. In this work, we gathered evidence indicating that the cAMP/Epac pathway is activated in different cells lines. In accordance, data collected from pull-down experiments designed to identify only the active form of Rap1b (Rap1b-GTP), and infection assays using cells transfected with a constitutively active mutant of Rap1b (Rap1b-G12V), strongly suggest the participation of Rap1b as mediator of the pathway. In addition to the activation of this small GTPase, fluorescence microscopy allowed us to demonstrate the relocalization of Rap1b to the entry site of the parasite. Moreover, phospho-mimetic and non-phosphorylable mutants of Rap1b were used to demonstrate a PKA-dependent antagonistic effect on the pathway, by phosphorylation of Rap1b, and potentially of Epac. Finally, Western Blot analysis was used to determine the involvement of the MEK/ERK signalling downstream of cAMP/Epac/Rap1b-mediated invasion.


Introduction
Smooth muscle and heart are the most important target organs for T. cruzi infection and persistence during the chronic phase of Chagas disease. Taking into account that Epac has a critical role in cAMP-mediated invasion and the regulation of various cAMP-dependent functions in smooth muscle and heart, possibly modulating the intracellular concentration of Ca 2+ through the activation of Rap1 and the participation of ERK1/2 [18,22,23], deciphering the detailed functioning of the cAMP/Epac pathway would provide a deeper insight into the host cell invasion mechanisms mediated by this cyclic nucleotide. In this work, we investigated the involvement of two known effectors, Rap1b and ERK, as potential mediators in the cAMP/ Epac-dependent invasion by T. cruzi and the role of PKA-dependent Rap1b phosphorylation.

Cells and parasites
NRK (ATCC CRL-6509), VERO (ATCC CCL-81) and HELA (ATCC CCL-2) cell lines were cultured in DMEM medium supplemented with Glutamax (Gibco), 10% (v/v) FBS (Natocor), 100 U/ml penicillin and 0.1 mg/ml streptomycin (Sigma), and maintained at 37˚C in a 5% CO 2 atmosphere. The HL-1 cell line [24] was cultured in a gelatin/fibronectin matrix (5 μg fibronectin / 0.02% gelatin (m/v)-Sigma) and Claycomb culture medium supplemented with Glutamax(Gibco), 10% (v/v) FBS, 100 U/ml penicillin, 0.1 mg/ml streptomycin and 0.1 mM norepinephrine (Sigma). Tissue culture-derived trypomastigotes forms (TCT) of T. cruzi Y strain were routinely maintained in VERO cells cultured in DMEM supplemented with 4% FBS and penicillin/streptomycin. Trypomastigotes were obtained from supernatants of infected VERO cells by centrifugation. First, the supernatant conditioned medium was centrifuged at low speed (500 g) to remove intact cells and cell debris Then, the supernatant obtained was centrifugated at 3,000 g for 15 min. and the pellet with the parasites was washed in PBS three times.

Invasion assay
Cells were grown on glass cover slides in a 24 multi-well plate with DMEM 10% FBS for 24 hours at 2x10 4 cells/well density at 37˚C, 5% CO 2 and incubated with: 37,5 μM of the Epac1 inhibitor ESI-09 (Sigma); 300 μM of 8-Br-cAMP (Biolog); 50 μM of the MEK1/2 inhbitor PD98059; or 0,1% DMSO as a control condition. Cells were then washed and infected with trypomastigotes of the Y strain (moi 100:1) for 2 hours. Parasite were removed and cells incubated for 48 hs. Cells were fixed, stained with DAPI and infection level determined by fluorescence microscopy. Percentage of infection (calculated as the (#Infected cells/total counted cells) � 100) and amastigotes/100 cells were calculated counting 3,000 cells, expressed as mean ± SD of three or more independent experiments and performed in triplicate. Infection of non-treated cells was considered as basal infection.

Host cell transfection
A transient transfection protocol with polyethyleneimine (PEI) was used [25]. Briefly, cells were grown at about 60% confluence and incubated at 37˚C in a 5% CO 2 , 95% humidified air environment. Next day, cells were transfected with pCGN empty vector (EMPTY), pCGN-HA-Rap1b (HA-Rap1) or HA-Rap1b mutants (G12V, S179A, S179D, or combinations) (kindly provided by Dr D. Altschuler, University of Pittsburgh, USA) using a ratio of 4:1 PEI: DNA mix in OptiMEM medium (Gibco). The mixture was kept for 30 min. at room temperature and then added to the cells and incubated at 37 C and 5% CO 2 . After 24h, cells were washed with PBS and complete medium (DMEM or Claycomb 10% FBS) was added. The transfected cells were used at 24h post-transfection.

Trypomastigote release assay
HL-1 cells were seeded on a 24-well plate at a concentration of 7000 cells/mL in Claycomb medium supplemented with 10% FBS. After 24 hours, cells were infected and treated as described above. 72 hours later, medium was replaced with fresh prepared treatments until trypomastigotes were observed under microscope at six days post infection (pi). Supernatants were transferred to a new plate, washed to avoid mammalian cell contamination and a solution of resazurin sodium salt was added as a fluorogenic oxidation-reduction indicator (final concentration 0.1 mM). After 3 hours of incubation, fluorescence was measured with a FLUOstar OPTIMA (BMG LABTECH) microplate reader at 590 nm (excitation: 570 nm). Baseline corrected values of fluorescence were normalized to the negative control. Results are expressed as mean ± SD of three or more independent experiments and performed in triplicate.

GST Pull-down
Detection of active Rap1 (GTP-bound) was performed through pull-down assays using a recombinant GST-RBD protein (GST fusion to the Rap1b-binding domain of the RalGDS protein, which only recognizes active Rap). A total of 1 mL bacteria lysates containing GST or GST-RBD were mixed by rotation with 40 μl 50% GSH-Sepharose at 4˚C for 1 h. The beads were centrifuged at 800 g for 2 min. at 4˚C and washed with lysis buffer. Lysates from HA-Rap1 transfected cells pre-treated for 2h with 8Br-cAMP, infected with trypomastigotes of the Y strain (Tp Y) or mock infected (Ctrl) were incubated with RBD-glutathione-agarose resin for 1h at 4˚C. Resin was washed and eluted with cracking buffer for WB analysis.

ERK phosphorylation
Cells were treated with or without PD98059 for 2h and incubated with trypomastigotes of the Y strain for 30 min., treated with 750 μM H 2 O 2 for 5 min. or mock infected. Then, cells were lysed and cracking buffer added for WB analysis.

Indirect immunofluorescence assay
Cells were adhered to glass previously treated with 40 μg/ml of poly-D-lysine (Sigma), fixed with PBS-PFA 4% (Sigma), washed with PBS and incubated with NH 4 Cl for 15 min. Then, were permeabilized with 0.2% Triton-x100 and incubated with anti-RAP1 antibody (Genscript) at 4˚C. After 16 h, washed with PBS, incubated with mouse anti-IgG (H+L) anti-conjugated to Alexa Fluor594 antibody (Jackson InmunoResearch), and nuclei stained with DAPI. Finally, glasses were mounted on slides with FluorSave mounting solution (Merk Millipore). Preparations were analysed in a Nikon Eclipse E600 fluorescence microscope.

cAMP/Epac activation as a ubiquitous mechanism of invasion in T. cruzi
The crucial role of Epac during invasion by T. cruzi was recently described in NRK cells [7]. In order to assess the ubiquity of the cAMP/Epac pathway, other cell lines were used in invasion assays. Similar to what happened in NRK cells [4,7], high levels of cAMP induced by a nonhydrolysable permeable analogue of cAMP, 8-Br-cAMP (8-Bromoadenosine 3 0 ,5 0 -cyclic monophosphate) (Biolog), positively modulated invasion in both HELA and HL-1 cells (Fig  1). Consistent with this result, specific pharmacological inhibition of Epac by ESI-09 (Sigma), resulted in a significant decrease in invasion in both cell lines (Fig 1).

PLOS NEGLECTED TROPICAL DISEASES
In accordance with these results, when HELA and HL-1 cells transfected with a constitutively active mutant of Rap1b, Rap1b-G12V, in which a single point substitution, glycine-tovaline at codon 12 of Rap1b, a significant increase in infection was observed when compared with the control (Fig 3A-3D, respectively). Noteworthy, when the complete invasion-differentiation-release cycle was evaluated in HL-1 cells overexpressing Rap1b-G12V (Fig 3E), trypomastigotes released into the medium showed similar results than the results obtained for percentage of infected cells and amastigotes/100cells ( Fig 3C and 3D), suggesting the cAMP/ Epac/Rap1b pathway would play a role in the early steps of the establishment of the infection, as previously hypothesized [7]. In addition to these results suggesting that Rap1b-GTP is required as a mediator of the cAMP/Epac1 pathway during the invasion by T. cruzi, fluorescence microscopy in HL-1 (Figs 4 and S3) and Hela cells (S4 Fig), revealed the relocalization of Rap1b, reflected as an increase in the fluorescence intensity of Rap1b, to the site of entry of T. cruzi, supporting the hypothesis that Rap1b needs to be activated and properly localized in the entry site.

Role of PKA-dependent Rap1b phosphorylation
While the specific activation of PKA had no effect on invasion, an increase in internalized parasites was observed as a result of PKA inhibition [7]. Therefore, under physiological conditions, PKA-mediated phosphorylation would negatively regulate the cAMP/Epac pathway of invasion. The inhibition of the Epac-mediated invasion pathway could be achieved, at least, at two different levels: through direct phosphorylation of Epac or at the level of Rap1, an Epac

PLOS NEGLECTED TROPICAL DISEASES
Host Cell Rap1b mediates T. cruzi invasion effector and a known target for PKA-mediated phosphorylation [26]. Since trypomastigotes released (Fig 3E), showed similar results than the percentage of infected cells and amastigotes/ 100cells (Fig 3C and 3D), we used this more reliable technique, in comparison to cell counting in the microscope, to evaluate the role of Rap1b phosphorylation. HL-1 cells overexpressing phospho-mimetic (S179D) or phospho-deficient (S179A) Rap1b mutants were infected and trypomastigotes released at day 6 pi were counted using resazurin method [27]. As shown in In the case of the trypomastigote release assay, HL-1 cells were infected and treated as described above. 72 hours later, medium was replaced with fresh prepared treatments until trypomastigotes were observed under microscope at six days post infection (pi). Supernatants were transferred to a new plate and quantification of trypomastigotes was performed with resazurin method. Results are expressed as mean ± SD (n � 3). ���� p<0.0001, t student test.
https://doi.org/10.1371/journal.pntd.0011191.g003 phospho-mimetic Rap1b-S179D mutant presented a decrease in the number of released trypomastigotes, with respect to control cells or cells overexpressing the non-phosphorylable mutant Rap1b-S179A, supporting a PKA-dependent antagonistic effect on the pathway, as previously described in NRK cells [7]. Interestingly, the effect of phosphorylation could be reverted by transfecting cells with the double mutant G12V/S179D, a constitutive active phospho-mimetic Rap1, opening the possibility of a two-level regulation of PKA on the Epac/Rap1 pathway.

MEK/ERK as a downstream effector of cAMP/Epac-mediated invasion of T. cruzi
In order to elucidate the involvement of MEK/ERK in the cAMP-dependent invasion, activation of ERK1/2 was analysed by Western Blot. An increase in ERK1/2 phosphorylation in both NRK ( Fig 6A)  To determine whether the activation of ERK1/2 modulates the infection levels, cells pretreated with the MEK1/2 kinase inhibitor PD98059 were infected with the parasite. In accordance, the inhibition of ERK phosphorylation produced a significant decrease in infection ( Fig  6B). MEK/ERK could be independently activated or a downstream effector of Epac/Rap1. The fact that the inhibition of MEK or Epac induced a similar decrease in invasion, and no additive or synergic effects were observed when both proteins were simultaneously inhibited (Fig 6C), suggests that MEK/ERK is a downstream effector of cAMP/Epac1/Rap1b-mediated invasion.

Discussion
Infection of cell cultures shown to be a useful model to study host-cell interaction and invasion [28], however, it has to be considered that T. cruzi invasion showed to be a complex process just taking into account the different stages of the parasite with the ability to infect host cells.

PLOS NEGLECTED TROPICAL DISEASES
In addition, despite of being able to infect any nucleated cell in vitro, it has been shown in vivo that T. cruzi exhibits a certain cellular tropism [29] and that the signalling pathways activated in the host cell differ according to the stage of the parasite [30]. This complexity is even higher when considering different DTUs, strains and the repertoire of surface/secreted molecules that activates different signalling pathways in the host cell [1]. Additionally, differences in cell recognition and invasion, when performing assays in different cell lines, should be also consider [31]. With all these contemplations, and when it is clear that the possibility of replicating human biology in a dish is quite limited, in vitro models are an essential tool to dissect the phenomenon of parasite-host cell interaction and invasion.
In this context, it has been reported that the activation of cAMP-mediated signalling pathways triggers Ca 2+ -dependent lysosomal exocytosis and promotes host cell invasion by T. cruzi [4]. The Ca 2+ release from intracellular compartments, such as the endoplasmic reticulum, is associated with an increase in intracellular levels of cAMP. In mammalian cells, cAMP downstream effectors, PKA and Epac, are involved in Ca 2+ -activated exocytosis events [32]. Furthermore, members of these pathways, including Rap1, have been localized to late endosomes/ lysosomes [15], and Epac-mediated activation of Rap1 has been identified in regulated exocytosis in human sperm [33], insulin secretion [34], and pancreatic amylase release [35]. It was previously shown that Epac1-mediated signalling represents the main mechanism for cAMPmediated invasion by T. cruzi [7]. In addition, ERM proteins, which are essential for the function and architecture of the cell cortex by linking the plasma membrane to the underlying actin cytoskeleton [11], have been associated with the invasion of EAs [12]. Moreover, in confocal studies, it was shown that ERM proteins are recruited at the entry site of the parasites where they colocalize with F-actin, while its depletion inhibits HELA cells invasion [12]. Remarkably, one of its members, radixin, was identified as a scaffold unit for cAMP effectors in the spatial regulation of Epac1/Rap1-mediated signalling [9,10]. In this regard, we have previously revealed a link between Epac1 and radixin in the cAMP-mediated invasion of TCTs, by blocking the invasion of NRK cells with a permeable peptide of 15 amino acids that binds to the minimal ERM-binding domain of Epac [7]. In order to elucidate the role of cAMP

PLOS NEGLECTED TROPICAL DISEASES
Host Cell Rap1b mediates T. cruzi invasion downstream effectors involved in T. cruzi invasion, we evaluated the activation of the cAMP/ Epac pathway by TCTs of Y strain in NRK, HELA and HL-1 cell lines. NRK cells are normal fibroblasts from rat kidney, originally used in the establishment of cAMP as a modulator of invasion events [4] and to demonstrate the participation of Epac1 as the main effector of this modulation [7]. On the other hand, HELA cells are epithelial human cervix cells that have been widely used in invasion assays [36][37][38] and HL-1 cells, previously used in invasion assays, as well [39], are cardiomyocytes from mouse heart, one of the most important target organs in the infection and persistence of T. cruzi. Our data showed that the activation of the cAMP/ Epac pathway by TCTs occurs regardless of the origin (rat, mouse, human) or the cell type (kidney, cervix, heart) that the parasite is invading. In addition, we investigated the role of Rap1b during the cAMP/Epac1-mediated invasion. Rap1b, a GTPase of the Ras family, is known to integrate Epac-and/or PKA-dependent events to achieve an efficient cAMP signal transduction [18,40,41]. Pull-down assays were used to detect higher levels of activated GTPbound Rap1 in lysates from infected cells. Likewise, as shown in Fig 3, cells transfected with the constitutively active form of Rap1b (G12V) were more susceptible to infection, compared to the control. However, it is important to note that due to the fact that cells are constitutively overexpressing Rap1b (G12V), and cells were incubated for 48 hs after invasion in order to achieve sensitive in the parasite count, an effect of the overexpression of this small GTPase on parasite replication could not be excluded, and needs to be further explored. In addition to Rap1b activation, and in accordance to our hypothesis, data obtained from fluorescence microscopy assays evidenced the recruitment of Rap1b to the parasite entry site. Interestingly, when studying PKA participation using a specific inhibitors of this kinase, it was observed that the invasion levels of TCTs increased compared to the control [7], suggesting a PKA-dependent antagonist effect. This effect could be mediated by PKA phosphorylation of the effectors of the cAMP pathway, such as Epac and Rap1b. PKA-dependent phosphorylation at S179 of Rap1b has long been established [42]. Results presented in this work support the antagonistic effect of PKA through, at least, Rap1b phosphorylation, since trypomastigote release was affected in cells transfected with phospho-mimetic Rap1b-S179D, with respect to control cells and cells overexpressing Rap1b-S179A, the non-phosphorylable version of Rap1b. In line with these observations, it has been shown that Rap1b phosphorylation destabilizes the association of this protein with the plasma membrane and promotes Rap1b inactivation [43,44]. Nevertheless, as mentioned above, the experiments were carried out in transfected cells, and an effect of the overexpressed proteins on parasite replication could not be ruled out.
Overall, our data suggest that activation and relocalization of Rap1b are required as a mediator of the cAMP/Epac1 pathway during the TCT infection. In this scenario, the fact that PKA negative regulation on infection was abrogated in the presence of the constitutively active G12V mutation, suggests that Rap1b is required in the phosphorylated and inactive form to completely abolish the cAMP/Epac/Rap1b pathway of infection.
It has been described that the MEK/ERK pathway can be activated or inhibited by cAMP [45]. Furthermore, the activation of this pathway participates in the invasion of T. cruzi by way of the interaction of the host cell with parasite surface molecules, such as TS [46], Tc85 [47] or TSSAII [48]. Also, Rap1 is associated with the phosphorylation and activation of ERK1/2 in smooth muscle [19]. Accordingly, our data revealed that TCTs induce ERK1/2 phosphorylation in mammalian cells and ERK1/2 activation modulates the invasion of these parasites as a downstream effector of Epac/Rap1-mediated invasion.
Although the transient increase in cytosolic Ca 2+ concentration and lysosome recruitment that occur during invasion are characteristics shared between MTs and TCTs [4,5], the signalling pathways triggered by both forms of parasites in the host cell are different. In TCT invasion, ERK1/2 activation is a distinctive feature that is mediated by Ca 2+ -dependent lysosomal exocytosis through the regulation of F-actin and the activation of the focal adhesion kinase (FAK) [49]. During MT invasion, in contrast, PKC promotes Ca 2+ release from inositol 3-phosphate (IP3)-sensitive compartments through the binding of the surface glycoprotein gp82 to LAMP-2 receptors [30,50]. On the contrary, the activity of PKC is not required for the invasion of TCTs in NRK cells, since treatment with PKC inhibitors did not affect the response to Ca 2+ or the reorganization of F-actin, and has no effect on parasite internalization [51]. The divergence between the signalling pathways triggered by MTs and TCTs might be associated with the fact that the internalization of TCT is initiated by an invagination of the plasma membrane [52], in a lysosomal exocytosis-dependent process induced by a membrane injury and the following activation of the PMR mechanism [53]. These mechanisms lead to changes that take place through the inhibition of the Rho/Rho signaling pathway by PKA [54,55]. The fact that RhoA promotes actin polymerization but has a negative effect on EAs internalization during HELA cell invasion [56] and that Rap1b inhibits RhoA/ROCK activity in the muscle smooth tissue [57], suggest the hypothesis that the cAMP/Epac1/Rap1b signalling pathway could be activated in the first steps of the invasion by T. cruzi, promoting Ca 2+ -dependent lysosomal exocytosis and the reorganization of the cytoskeleton. Once the parasite is inside the cell, a PKA-mediated inhibition of Epac/Rap1b might be necessary for the parasite retention. In accordance, our results showed that Rap1b seems to be associated with the plasma membrane at the parasite entry site where it could be required during the internalization process and PKA had an antagonistic effect, probably through the phosphorylation of the S179 of Rap1b.
In this work, we have gathered evidence strongly suggesting that the cAMP/Epac/Rap1b/ ERK pathway is activated during the early steps of host cell infection and that it would be negatively regulated by PKA, possibly through the phosphorylation of Epac and/or Rap1b. Importantly, a detailed characterization of effectors involved in T. cruzi invasion would provide an attractive set of new therapeutic targets for the repositioning or the development of new antiparasitic drugs, since there is a large variety of therapies that target cAMP-mediated signalling [58].

S1 Fig. Densitometry analysis.
Chemiluminescence was recorded with the C-DiGit scanner (LI-COR), and bands were quantified and normalized against the input using ImageJ and Ima-geLab 6.1 (Bio-Rad) software. The normalization was performed following the "Western Blot Normalization Using Image Lab Software" guide. Results are expressed as mean ± SD (n�3). � p<0.05, �� p <0.005, One-way ANOVA-Dunnett´s multiple comparison test. (PDF) S2 Fig. Rap1b pull-down assays. A) HA-Rap1 transfected HELA cells were incubated for 2 h 37.5uM ESI-09 or 0.1% DMSO. Then, cells were lysed and pull-down assay with glutathioneagarose resin performed for 1 h at 4˚C. Resin was washed and eluted with cracking buffer for WB analysis. B) Bands were quantified and normalized against the input using ImageJ cell software. Results are expressed as mean ± SD (n�3). �� p<0.01, t student test.