Cell invasion by intracellular parasites – the many roads to infection

ABSTRACT Intracellular parasites from the genera Toxoplasma, Plasmodium, Trypanosoma, Leishmania and from the phylum Microsporidia are, respectively, the causative agents of toxoplasmosis, malaria, Chagas disease, leishmaniasis and microsporidiosis, illnesses that kill millions of people around the globe. Crossing the host cell plasma membrane (PM) is an obstacle these parasites must overcome to establish themselves intracellularly and so cause diseases. The mechanisms of cell invasion are quite diverse and include (1) formation of moving junctions that drive parasites into host cells, as for the protozoans Toxoplasma gondii and Plasmodium spp., (2) subversion of endocytic pathways used by the host cell to repair PM, as for Trypanosoma cruzi and Leishmania, (3) induction of phagocytosis as for Leishmania or (4) endocytosis of parasites induced by specialized structures, such as the polar tubes present in microsporidian species. Understanding the early steps of cell entry is essential for the development of vaccines and drugs for the prevention or treatment of these diseases, and thus enormous research efforts have been made to unveil their underlying biological mechanisms. This Review will focus on these mechanisms and the factors involved, with an emphasis on the recent insights into the cell biology of invasion by these pathogens. Summary: Intracellular parasites have evolved different ways to cross the barrier imposed by host cell plasma membrane and cytoskeleton. This Review discusses these diverse entry mechanisms from cell attachment to invasion.


Introduction
Evolving a way to invade their host cells is the first obligatory challenge for intracellular parasites. In order to establish intracellular life, these microorganisms employ numerous strategies to overcome the barrier that is imposed by the host cell plasma membrane (PM) and cytoskeleton. Small intracellular pathogens, such as bacteria and viruses, can be endocytosed through natural and abundant PM invaginations, via clathrin-, caveolae-and flotillin-mediated endocytosis (Cossart and Helenius, 2014). However, these small, nanoscale membrane invaginations cannot carry larger intracellular parasites, such as protozoans. The internalization of these pathogens can be achieved instead through several other routes. Larger pathogens can actively drive cell entry through the formation of a specialized cell machinery, such as the moving junction that is present in the apicomplexans Plasmodium spp. (Aikawa et al., 1978) and Toxoplasma gondii (Mordue et al., 1999). Invasion can also occur through the subversion of physiological host cell processes, such as lysosome-triggered Ca 2+ -dependent endocytosis, which is used by all nucleated cells to repair wounded PMs (Corrotte and Castro-Gomes, 2019); this pathway has previously shown to occur for Trypanosoma cruzi (Rodríguez et al., 1996;Fernandes et al., 2011), and also more recently by our lab for Leishmania amazonensis (Cavalcante-Costa et al., 2019). Large parasites can be equally phagocytosed, if they are able to resist phagocytic degradation and live within phagocytic cells, such as protozoans of the genus Leishmania (Zenian et al., 1979). Finally, internalization can also be induced through the formation of a specialized structure that directs the parasite towards the PM, as for microsporidians, for example, Encephalitozoon spp. (Xu and Weiss, 2005). Understanding the initial interaction between a host cell and an intracellular parasite is essential, as the molecular determinants involved could be exploited as vaccine targets to help block the infection from the onset. This Review will focus on the cellular and molecular determinants that mediate host cell invasion by the above parasites paying special attention to the initial steps.
Apicomplexans -Plasmodium spp. and Toxoplasma gondii Apicomplexans are obligate intracellular parasites characterized by the presence of an apical complex, a cellular structure crucial for cell invasion composed of cytoskeletal components and three secretory organelles, namely, micronemes, rhoptries and dense granules (Adl et al., 2018). They also lack flagella, cilia or pseudopods, thus relying on a specialized intracellular machinery that allows them to move when attached to substrates or the host cell PM. Toxoplasma gondii (agent of toxoplasmosis) and Plasmodium spp. (agents of malaria) are included within this phylum ( Fig. 1; Box 1).

Cell invasion by apicomplexans
Owing to their inability to undergo cell division and limited viability in extracellular environments, apicomplexans have evolved a unique and extremely effective mechanism for invading a host cell, which is dependent on gliding motility and coordinated secretion of proteins contained in their apical secretory organelles (Box 2) . In the initial moments of interaction with the host cell, the apicomplexan zoite (see Glossary) moves laterally on its surface until it encounters an ideal receptor that triggers host cell invasion; this then leads to an apical re-orientation of the zoite, which is mediated by microneme adhesins and is followed by the secretion of rhoptry proteins into host cell cytoplasm through a transient pore in the host cell PM (Carruthers and Boothroyd, 2007;Weiss et al., 2015;Dubremetz, 2007) (see Fig. 4A).
In T. gondii, the initial lateral weak interaction with host cell membrane is mediated by glycosylphosphatidylinositol (GPI)anchored proteins of the surface antigen glycoprotein (SAG)-related sequence (SRS) superfamily, which are resident on the parasite PM (Dzierszinski et al., 2000;Manger et al., 1998;Wasmuth et al., 2012). SRS genes are differently expressed according to lineage or evolutionary stage, and are involved in a wide range of functions, Left, an electron micrograph of a merozoite of P. knowlesi at the initial contact between the merozoite's apical end (arrow) and an erythrocyte (E). The merozoite shows an apical end (labeled A), a rhoptry (R), a nucleus (N) and a mitochondrion (M). The surface is covered with a surface coat (double arrow). Middle, the erythrocyte membrane is thickened (15 nm) at the attachment site (arrow). Inset, higher magnification micrograph of the erythrocyte-merozoite attachment site showing the thickened erythrocyte membrane. Right, an advanced stage of erythrocyte (E) entry by a merozoite (Mz). Note a junctional attachment (labeled C) at each side of the entry orifice. Republished with permission of Rockefeller University Press from Aikawa et al., 1978; permission conveyed through the Copyright Clearance Center. (B) Schematic representation of cell binding and formation of the moving junction during host cell invasion by the apicomplexan parasite Toxoplasma gondii.
(1) Rhoptry secretion. After re-orientation, the parasite apical complex faces the host cell PM. A local loosening of the host cell actin cytoskeleton is induced by the injection of parasite profilin, and rhoptry proteins are discharged into host cell cytoplasm.
(2) Moving junction formation (see also enlargement and electron micrograph underneath). Proteins secreted by the parasite and host cell proteins form a multi-molecular complex beneath host cell PM. This is seen in the electron micrograph (É.S.M.-D.) as an electron-dense region (black arrow) in a host cell (HC; here, a LLC-MK2 epithelial cell) being invaded by a T. gondii (Tg) tachyzoite. This is the region where the moving junction is formed. The enlargement shows a schematic representation of the moving junction. Rhoptry proteins (RON2, RON4, RON5 and RON8) form a molecular complex with host cell proteins of the ESCRT family (ALIX, TSG101, CIN85 and CD2AP); this anchors the parasite to the host cell actin cytoskeleton, while the extracellular domain of RON2 binds to AMA1 present on parasite PM. Inside the parasite, the cytosolic domain of AMA1 binds to a parasite actin-myosin motor through the glideosome-associated connector (GAC), while the rhomboid protease ROM4 successively disengages the complex to allow parasite moving. IMC, inner membrane complex. See also Box 2.
including cell adhesion (Dzierszinski et al., 2000;Kim et al., 2007;Leal-Sena et al., 2018;Tomita et al., 2018;Wasmuth et al., 2012). Interaction of SRS proteins with the host cell or substrates is, however, weak, possibly allowing the parasite to move laterally along their surface (Carruthers and Boothroyd, 2007). Disruption of the SAG3 gene, which codes for the SRS adhesive protein, reduces parasite adhesion to host cell and also decreases virulence of T. gondii to mice (Dzierszinski et al., 2000). It has also been shown that both SAG3 and SAG2 interact with heparin sulfate proteoglycans (Jacquet et al., 2001;Zhang et al., 2019). SAG1 also interacts with sulfated proteoglycans, as supported by structural studies and cell binding assays (He et al., 2002;Azzouz et al., 2013). The initial interaction between Plasmodium zoites and host cells also rely on parasite GPI-anchored proteins (Beeson et al., 2016), such as the circumsporozoite protein (CSP) (Yoshida et al., 1980) and merozoite (see Glossary) surface proteins (MSPs) (reviewed by Beeson et al., 2016). CSP is highly conserved among Plasmodium species and mediates sporozoite (see Glossary) migration from mammalian dermis to liver (Tewari et al., 2002). Upon reaching the liver, sporozoites switch from a migratory to an invasive mode. Sporozoite binding to heparan sulfate proteoglycans (HSPGs) that are expressed by hepatic stellate cells culminates in the secretion of cysteine proteases by the parasite (Coppi et al., 2005;Frevert et al., 1993;Pinzon-Ortiz et al., 2001;Zhao et al., 2016). This leads to a proteolytic processing of CSP, exposing its thrombospondin type-1 repeat (TSR) domain, leading to an increase in parasite adhesion, which allows invasion (Coppi et al., 2005). The weak interaction of merozoites with erythrocytes is mediated by MSPs (Gilson et al., 2006). MSP1 is the most abundant and well-studied MSP (Cowman and Crabb, 2006), and is considered an important target for vaccine development (Beeson et al., 2016). MSP1 is proteolytically processed by the subtilisin-like protease 1 into four fragments (denoted p42, p38, p30 and p83) (Koussis et al., 2009), which have been implicated in the binding to the erythrocyte surface proteins band 3 (also known as SLC4A1) and glycophorin A (Baldwin et al., Box 1. Transmission and parasite biology of T. gondii, Plasmodium spp., T. cruzi, Leishmania spp. and microsporidians T. gondii is transmitted through ingestion of sporozoite-containing oocysts from felids or encysted bradyzoites present in infected undercooked meat. After ingestion, sporozoites or bradyzoites infect gastrointestinal epithelial cells in the ileum (Dubey, 2007;Dubey et al., 1997;Speer and Dubey, 1998). Bradyzoites or sporozoites then differentiate into tachyzoites, which spread throughout the body, generating pathology (Dubey et al., 1997;Speer and Dubey, 1998).
Infection by Plasmodium spp. occurs through the inoculation of sporozoites into the skin of vertebrate hosts by anopheline mosquitoes (Matsuoka et al., 2002;Sidjanski and Vanderberg, 1997). Sporozoites then migrate to the liver and infect hepatocytes, in which they replicate (reviewed by Yang and Boddey, 2017), originating merozoites. After reaching the bloodstream, merozoites invade erythrocytes, in which they are able to replicate and, thus, generate additional merozoites that can be released and cyclically invade new erythrocytes, thereby giving rise to the erythocytic cycle of malaria.
T. cruzi is transmitted to humans as flagellated metacyclic trypomastigotes by kissing-bug insects of the Reduviidae family (Coura and Dias, 2009;Araújo et al., 2009). These forms evolve in the invertebrate gut from replicative forms, the epimastigotes. After blood repast, the vector feces containing trypomastigotes come into contact with either the mucosa or skin micro-lesions at the bite site. Trypomastigotes can then invade any nucleated cell, where they transform into replicative oval-shaped forms with unapparent flagella, the amastigotes, which differentiate again into trypomastigotes. These forms can invade neighboring cells or reach the bloodstream, thereby either spreading to other tissues or being ingested by a new insect vector (Brener, 1973).
Leishmania spp. are transmitted by the bite of a sand fly vector, which inoculates flagellated infective parasites, the metacyclic promastigotes, into the dermis of vertebrate hosts (Handman, 1999). Promastigotes can transiently invade different cell types (Peters et al., 2008;Williams, 1988;Kautz-Neu et al., 2012;Moll et al., 1993;Bogdan et al., 2000;Cavalcante-Costa et al., 2019), before reaching macrophages, their final destination. Inside macrophages, they nest within vacuoles, where they transform into replicative oval-shaped amastigote forms, causing different clinical manifestations of the disease, depending on the species of Leishmania and on host immune responses.
Infection by microsporidians is established through the inhalation or ingestion of virulent spores, which can infect several cell types (reviewed by Han and Weiss, 2017). Upon contact with host cells, the spores extrude a long infection apparatus, the polar tube, which drives the parasite into the host cell. Parasites are internalized in a PV, where it replicates, generating more spores, which leave the host cell to infect neighboring cells, thereby amplifying infection (CDC -DPDx -Microsporidiosis; https://www.cdc.gov/ dpdx/microsporidiosis/index.html).
Box 2. Apicomplexansmicroneme and rhoptry secretion and gliding motility during host cell invasion Microneme protein secretion in T. gondii and Plasmodium involves Ca 2+ signaling (Dawn et al., 2014;Wetzel et al., 2004) that is initiated by its mobilization from intracellular stores by cGMP (T. gondii) (Bullen et al., 2016) or cAMP (Plasmodium) (Dawn et al., 2014) signaling. Ca 2+ increase in the cytosol activates Ca 2+ -dependent protein kinases, regulating microneme secretion (Lourido et al., 2010;Wetzel et al., 2004;Singh et al., 2010;Wernimont et al., 2010) as well as microneme and PM fusion, which is mediated by the parasite protein DOC2.1 protein (Farrell et al., 2012). In T. gondii, phosphatidic acid also participates in this secretion process by binding to the acylated pleckstrin homology domain-containing protein (APH) that is present in micronemes (Bullen et al., 2016). Rhoptry secretion is followed by strong binding of the parasite to the host cell surface, which is mediated by specific microneme-secreted adhesins (Kessler et al., 2008;Singh et al., 2010). In T. gondii, MIC8 appears to participate in rhoptry secretion (Kessler et al., 2008), while in Plasmodium, the interaction of the microneme adhesin EBA175 with its receptor glycophorin A decreases cytosolic Ca 2+ levels in merozoites and triggers rhoptry secretion (Singh et al., 2010). A cytoplasmic protein homologous to the ferlin family of Ca 2+ -sensing proteins (TgFER2) has been shown to be essential in rhoptry secretion in T. gondii and points to a role for Ca 2+ signaling in the secretion of this organelle during invasion (Coleman et al., 2018). In addition, a newly identified set of proteins termed rhoptry apical surface proteins (RASPs) appear to be essential for rhoptry secretion in both T. gondii and P. falciparum by promoting the rhoptry fusion with parasite PM in a Ca 2+ -independent manner (Suarez et al., 2019).
Gliding is a substrate-dependent motility that is mediated by a protein complex, called the glideosome (Opitz and Soldati, 2002), which has an actomyosin motor that is composed of myosin A, myosin light chain, myosin essential chain and actin filaments (Boucher and Bosch, 2015). Motility is generated after the displacement of the myosin A head from the actin filament, which occurs when the extracellular domain of a transmembrane micronemal protein (TRAP in Plasmodium and MIC2 in T. gondii) strongly interacts with a receptor at the host cell surface or in the extracellular matrix. A physical interaction between the cytoplasmic tail of TRAP or MIC2 and the actin filament is mediated by a conserved protein, named glideosome-associated connector (GAC) (Jacot et al., 2016). Thus, gliding is the result of the displacement of myosin A from actin filaments, followed by the release of the parasite from the host cell, after the proteolytic removal of TRAP or MIC2 by an intramembrane serine protease of the rhomboid family protease (ROM4) (reviewed by Frénal et al., 2017), which then propels the parasite body forward.
The key factors in the recognition and strong adhesion of apicomplexan to the host cell surface are proteins secreted by the micronemes (and also by rhoptries in Plasmodium), the microneme adhesive proteins (MICs). Micronemal proteins are the first to be secreted (Box 2) and are essential for both host cell adhesion and parasite motility and invasion. Secreted MICs are not resident in the PM, but the presence of hydrophobic domains on some MICs allows an individual MIC or their complexes to be embedded in the parasite membrane (Huynh et al., 2003;Meissner et al., 2002;Reiss et al., 2001;Sheiner et al., 2010). MICs contain either a single adhesive domain or a variety of combinations. Among the adhesive domains found in T. gondii MICs are the microneme repeat domain (MAR family) (Blumenschein et al., 2007), the type A and TSR domains (found in the TRAP family) (Wan et al., 1997) and Apple/ PAN domains , as well as the chitin-binding-like (CBL) domain (Cérede et al., 2002). Whereas some domains, such as TSR, are conserved among many species of Apicomplexa, others are more restricted, such as Apple and MAR, which are specific for coccidians . Binding assays have shown that the MAR domain in T. gondii (Tg)MIC1 and TgMIC13 selectively interacts with sialylated oligosaccharides, which are common terminal carbohydrates of the glycocalyx in vertebrate cells . In addition, the TSR domain of TgMIC2 is important for T. gondii motility and cell invasion (Huynh and Carruthers, 2006). Indeed, the interaction of TgMIC2 type A domain with intercellular adhesion molecule 1 (ICAM-1) is possibly involved in parasite migration across polarized epithelial cells (Barragan et al., 2005).
In Plasmodium sporozoites, the micronemal proteins P36 and P52 of 6-cysteine domain proteins (the s48/s45 family) are essential for hepatocyte invasion (Ishino et al., 2005). P52 is a membrane GPI-anchored protein and acts as scaffold for P36 (Arredondo et al., 2018). This scaffold is then responsible for the interaction with different receptors at the hepatocyte surface, such as CD81 and scavenger receptor BI (SR-BI, also known as SCARB1) for P. falciparum and P. vivax sporozoites, respectively (Manzoni et al., 2017). Regarding merozoites, here, proteins belonging to the Duffy binding-like family (DBL), reticulocyte binding-like (RBL) family and TRAP family (containing TSR and type A domains) (Boucher and Bosch, 2015;Beeson et al., 2016) are responsible for erythrocyte invasion. Microneme DBLs (only found in Plasmodium spp.) recognize different receptors, such as sialylated glycoproteins (e.g. glycophorins) by erythrocyte-binding antigens (EBAs) (Cowman et al., 2017) or specific proteins, as for the binding of the DBP in P. vivax to Duffy antigen (also known as ACKR1) located on red blood cells (Adams et al., 1992;Batchelor et al., 2014). Proteins of the rhoptry RBL family in P. falciparum mediate a sialic acid-independent invasion; for example, Rh4 binds to complement receptor 1 (CR1) and Rh5 to basigin, which are both present on the erythrocyte PM (Crosnier et al., 2011;Tham et al., 2010). During invasion, P. falciparum (Pf )Rh5 forms a complex with two other proteins, Rh5-interacting protein (PfRipr) and cysteine-rich protective antigen (CyRPA) (Chen et al., 2011;Dreyer et al., 2012). This possibly mediates the formation of a pore through the membranes of both merozoite and erythrocyte, which allows the release of Ca 2+ or the injection of parasite proteins inside erythrocytes (Volz et al., 2016;Weiss et al., 2015). Owing to its essential role in basigin binding during erythrocyte invasion by merozoites, Rh5 is a leading candidate antigen for vaccine development against P. falciparum infection (Payne et al., 2017). Microneme TRAP family members, such as those relevant for T. gondii, also have a role in Plasmodium spp. zoite motility and are important for skin-to-liver migration and cell invasion (Morahan et al., 2009;Moreira et al., 2008;Sultan et al., 1997). Compared to T. gondii, TRAP members in Plasmodium spp. are more diverse, and specific TRAP proteins are expressed in the different zoite stages (Morahan et al., 2009). A recent study has shown that the human cell-surface integrin αvβ3 could act as a receptor for TRAP in sporozoites (Dundas et al., 2018). In merozoites, MTRAP binds to semaphorin-7A at the erythrocyte surface (Bartholdson et al., 2012), and is important not only for motility and invasion, but also for gamete egress from erythrocytes inside the insect vector (Bargieri et al., 2016;Baum et al., 2006).

The formation of the moving junction
Rhoptry proteins are subcompartmentalized into two distinct regions with different functions during the invasion process (Counihan et al., 2013); neck proteins act in the formation of a structure called the moving junction (MJ) or tight junction , whereas proteins from the bulb have a broad spectrum of functions, such as in cell adhesion (Beeson et al., 2016), parasitophorous vacuole (PV) formation (Ghosh et al., 2017) and immune evasion (Hakimi et al., 2017). The existence of a MJ was first described by electron microscopy observations of Plasmodium knowlesi and is morphologically characterized by the close apposition between the parasite and host cell PMs, associated with an electron-dense region right below the host cell PM (Aikawa et al., 1978) (Fig. 1A). MJs are conserved in T. gondii (Fig. 1B) and Plasmodium spp., and their main molecular components are the rhoptry neck proteins (RON)2, RON4 and RON5, as well as the microneme protein apical membrane antigen 1 (AMA1), a transmembrane protein secreted at the apical cap of the zoite Glossary Zoite: a general name for the infective forms of Apicomplexan parasites that invade cells.
Oocyst: infective form of T. gondii delivered in feline feces (which contains infective sporozoites) that can be ingested by the mammalian host. Bradyzoite: infective and resistant form of T. gondii found in host-tissue cysts that can be ingested by the mammalian host. Tachyzoite: infective and replicative form of T. gondii responsible for parasite spread throughout different host tissues. Sporozoite: infective form of Plasmodium spp.; it is inoculated by the insect vector in the host dermis. Sporozoites travel along host body bloodstream and invade hepatocytes in the liver originating the exoerythrocytic phase of malaria. Merozoite: infective form of Plasmodium spp.; it is first delivered by the infected hepatocyte, reaches the bloodstream and then invades host erythrocytes. After replication inside erythrocytes, more merozoites are cyclically delivered, giving rise to the erythrocytic phase of malaria. Trypomastigote: infective form of T. cruzi present in the feces of the insect vector; it is deposited in host skin during vector blood meal. This form is also found in the mammalian host bloodstream after parasite replication as the amastigote form and is responsible for parasite spread throughout different host tissues. Amastigote: intracellular replicative form of T. cruzi or Leishmania spp. found in host cell cytosol or whiting vacuoles, respectively. Epimastigote: extracellular replicative form of T. cruzi found in the intestine of the insect vector. This form originates the trypomastigote form within the insect vector. Promastigote: infective form of Leishmania spp.; it is inoculated by the insect vector in host dermis during vector blood meal. Lebrun et al., 2005;Bradley et al., 2005;Besteiro et al., 2009;Curtidor et al., 2011;Narum et al., 2008). Additionally, RON4 L1 and RON8 are found in T. gondii MJs (Guérin et al., 2017;Straub et al., 2009), and RON2 is inserted into the host PM, acting as a receptor for AMA1, which is located on the parasite apex (Besteiro et al., 2009;Cao et al., 2009;Lamarque et al., 2011;Richard et al., 2010). Thus, zoites provide their own interacting ligand on the host cell to mediate the invasion process. The interaction between RON2 and AMA1 is of high affinity; thus, once the MJ is formed, zoites are committed to invasion (Delgadillo et al., 2016;Srinivasan et al., 2011), which then occurs within seconds (Morisaki et al., 1995). Whereas RON2 and AMA1 are localized on the interface between the zoite and host cell surfaces, RON4, RON4 L1 , RON5 and RON8 are localized in the host cytoplasm (Beck et al., 2014;Guérin et al., 2017;Straub et al., 2011). During MJ assembly, cytosolic RONs in T. gondii cooperate in recruiting host adaptor proteins, such as CIN85 (also known as SH3KBP1), CD2AP and the ESCRT-I components ALIX (also known as PDCD6IP) and TSG101, which might facilitate the physical interaction between the RON complex and the cortical actin cytoskeleton (Guérin et al., 2017) (Fig. 1B). Furthermore, secretion of toxofilin by T. gondii during invasion leads to disassembly of the actin meshwork at the site of entrance, permitting anchorage of the RON complex to a newly formed, ring-shaped F-actin structure in the MJ region (Gonzalez et al., 2009;Delorme-Walker et al., 2012). Such a ring-shaped F actin structure was also observed at the site of MJ during Plasmodium sporozoite invasion (Gonzalez et al., 2009). By contrast, an actin reorganization in erythrocytes linked to the MJ has not been observed during merozite invasion, but surprisingly, the presence of the cytoskeletal linker adducin was detected (Zuccala et al., 2016). The association of cytosolic RON proteins with the cortical cytoskeleton at the site of zoite entrance allows the MJ-anchored zoite to use the traction force of its actomyosin motor to make an invagination in the host PM (Bichet et al., 2014;Gonzalez et al., 2009;Straub et al., 2011). During T. gondii invasion, the MJ selectively excludes host transmembrane proteins from the forming PV membrane; this might prevent its fusion with the host lysosomes (Bichet et al., 2014;Charron and Sibley, 2004;Mordue et al., 1999), which would be deleterious to the parasite. Furthermore, during T. gondii invasion, disassembly of the actin meshwork underlying the PM, which is caused by the secretion of profilin during the initial host-pathogen interaction, also acts in decreasing the resistance of host cell membrane invagination (Delorme-Walker et al., 2012). Similarly, a cytoskeleton reorganization at erythrocyte entry site is also required for P. falciparum invasion (Dasgupta et al., 2014). This process is regulated by Ca 2+ signaling and phosphorylation of several erythrocyte cytoskeletal proteins, including β-spectrin, PIEZO1 and band 3, leading to a destabilization of the cytoskeleton (Fernandez-Pol et al., 2013;Wernimont et al., 2010;Zuccala et al., 2016). After invasion, the newly formed PV pinches off from the PM through an as-yet elusive mechanism. Experiments using the dynamin inhibitor dynasore in T. gondii suggest that the host cell dynamin participates in the pinching off of the PV (Caldas et al., 2009). However, more recently, it has been shown that the pinch off of the T. gondii PV is independent of the host cell, and appears to be induced by a twisting motion of the parasite, which mechanically mediates membrane scission and PV detaching (Pavlou et al., 2018).
Kinetoplastids are flagellated protozoans characterized by the presence of a large and unique mitochondrion. Close to the flagellar bag, this mitochondrion presents a condensation of its DNA, named the kinetoplast, which can be easily visualized in stained cells. The order Kinetoplastida is composed of several parasite protists of different genera that infect several animals, including humans. Important pathogenic species of this order are Trypanosoma cruzi (the agent of Chagas disease) and different species of the genus Leishmania (agents of leishmaniasis) ( Fig. 2; Box 1).

Cell invasion by Trypanosoma cruzi
Infection by T. cruzi starts with the binding of infective trypomastigote forms (see Glossary) to the host cell PM, which induces signaling pathways that culminate in its internalization into an endocytic vacuole (reviewed by Andrade and Andrews, 2004;Fernandes and Andrews, 2012). Early studies of T. cruzi-host-cell interaction demonstrated that, soon after invasion, T. cruzi resides in an acidic vacuole that contains lysosomal markers (Milder and Kloetzel, 1980;de Carvalho and de Souza, 1989). However, because invasion is not only independent of actin filaments (Schenkman et al., 1991), but also even facilitated by their disruption (Schenkman et al., 1991), phagocytosis can be excluded as a mechanism of parasite entry into host cells. This corroborates the fact that T. cruzi is able to invade any nucleated cell, including non-professional phagocytic cells ( Fig. 2A,B), such as myocytes, which are actually the main target cells of the parasite during chronic infection in the vertebrate host (Calvet et al., 2012).
The attachment of T. cruzi provokes a rise in Ca 2+ in the host cell cytosol, which signals the recruitment of host cell lysosomes to the parasite attachment site (Rodríguez et al., 1996) (Figs 2A, 4B). Lysosomes then fuse with the host cell PM, releasing their content to the external milieu. In doing so, they provide the membrane for the nascent PV, leading to parasite internalization. After complete enveloping of the parasite, the PV, which is rich in lysosomalassociated membrane proteins (LAMPs), is pinched off into the cytoplasm through as-yet-unidentified mechanisms. The source of the increased Ca 2+ in the host cell cytosol is either intracellular stores, mobilized by signaling triggered by surface and secreted proteins originating from the parasite, or from the extracellular milieu, which can influx through parasite-induced micro-injuries inflicted on the host cell PM (Rodríguez et al., 1997;Rodríguez et al., 1999;Fernandes et al., 2011). The understanding of this entry mechanism came from studies of a basic physiological process used by nucleated cells to repair PM wounds. Plasma membrane repair (PMR) is driven by (1) Ca 2+ influx, (2) lysosomal exocytosis and (3) extensive Ca 2+ -dependent and actin-independent endocytosis (Bi et al., 1995;Togo et al., 2000;Reddy et al., 2001;Idone et al., 2008). During PMR, one of the enzymes secreted by the lysosomes, acid sphingomyelinase (ASM), is able to trigger endocytosis through the formation of ceramide at the cell surface, which facilitates inward budding and thus endocytosis of the wounded membrane (McIntosh et al., 1992;Brown and London, 2000;Tam et al., 2010). Based on these findings, Fernandes and co-workers proposed that T. cruzi subverts PMR for entry into the host cell (Fernandes et al., 2011). Further evidence of PMR involvement in T. cruzi invasion comes from the observation that recently internalized parasites are found in ceramide-rich vacuoles that are formed by the action of ASM during parasite-induced PMR (Fernandes et al., 2011).
Several parasite surface proteins involved in invasion have been described (Ruiz et al., 1998;Caler et al., 1998;Scharfstein et al., 2000;Neira et al., 2003;Kojin et al., 2016;Alves and Colli, 2008; reviewed by Maeda et al., 2012). Among them are the members of the gp85/trans-sialidase (TS) super-family (gp82, gp90, gp85/TS, gp30); of these, gp82 has been recently shown to be recognized by the host-cell lysosomal protein LAMP-2, which may be present at low levels on the host-cell surface (Rodrigues et al., 2019). Binding of gp82 to target cells induces lysosome spreading and exocytosis, culminating in parasite internalization (Martins et al., 2011;Cortez et al., 2016). Two parasite enzymes have also been implicated in the induction of Ca 2+ signaling in host cells. Oligopeptidase B appears to generate a product that is recognized by a G-protein-coupled receptor (Tardieux et al., 1994;Burleigh and Andrews, 1995), while the cysteine protease cruzipain cleaves host kininogen into bradykinin, which interacts with its classical bradykinin receptor on the host (Scharfstein et al., 2000). Both stimuli eventually converge in a pathway that involves Ca 2+ signaling-mediated events and host cell lysosome recruitment, resulting in parasite internalization (Rodríguez et al., 1999;Caler et al., 1998). T. cruzi trypomastigotes can also interact with host extracellular matrix components, which appears to facilitate their contact with and invasion into host cells (Santana et al., 1997;Alves and Colli, 2008;Nde et al., 2012).
It is important to mention that amastigotes (see Glossary), which can be released extracellularly are also able to invade host cells (Fernandes et al., 2013). In contrast to trypomastigotes, however, invasion of amastigotes into host cells occurs through an actin polymerization-dependent pathway (Fernandes et al., 2013), and amastigotes are also able to induce phagocytosis in nonprofessional phagocytic cells .

Cell invasion by Leishmania spp.
Histological tissue sections of mammalians hosts with leishmaniasis reveal that macrophages are the ultimate host cells for Leishmania spp. (Benchimol and de Souza, 1981;Berman et al., 1981;Brazil, 1984;Davies et al., 1988). However, the circumstances under which macrophages are invaded in vivo by promastigotes (see Glossary) are not completely understood, particularly with regard to the molecular details. Nevertheless, in vitro, phagocytosis has been suggested as the main route of promastigote invasion, because, in the absence of host cell actin polymerization, macrophage infection is impaired (Akiyama and Haight, 1971;Alexander, 1975;Ardehali et al., 1979;Zenian et al., 1979;Roy et al., 2014) (Fig. 4C).
In vitro, phagocytosis of promastigotes by macrophages (Fig. 2D) appears to start only after two minutes of contact with parasites (Aikawa et al., 1982). It is remarkable that, during the first moments of contact, 90% of promastigotes are attached, with low affinity, to macrophages (Uezato et al., 2005) through their flagellar tip (Aikawa et al., 1982), suggesting a role for this structure in triggering phagocytosis. It is worth pointing out that promastigotes generally move in the direction of the flagellum, an anterior structure (Krüger and Engstler, 2015), that has been proposed to have a role in sensing the environment (Rotureau et al., 2009), probably as it is the first structure to encounter a host cell. After five minutes, however, parasites are tightly bound to macrophages and phagocytosed with no preferred orientation (Aikawa et al., 1982;Uezato et al., 2005). The vast majority of phagocytosed promastigotes clearly localize inside phagosomes (Courret et al., 2002), which promptly fuse with lysosomes (James et al., 2006;Courret et al., 2002); this creates an appropriate milieu for their fully differentiation into amastigotes (see Glossary), and subsequent replication (Alexander and Vickerman, 1975;Chang and Dwyer, 1976;Moradin and Descoteaux, 2012). Several macrophage receptors have been described to mediate parasite adherence and subsequent phagocytosis signaling (Mosser and Rosenthal, 1993;Lefevre et al., 2013;Polando et al., 2018Polando et al., , 2018Stafford et al., 2002;reviewed by Naderer et al., 2005 andWilson, 2012;Chauhan et al., 2017;von Stebut and Tenzer, 2018). One such type of receptor, the endocytic mannose receptors (MRs) in macrophages, which are pattern recognition receptors (PRRs), were previously considered to be necessary for attachment and phagocytosis of L. donovani and L. infantum promastigotes by macrophages (Blackwell, 1985;Wilson and Pearson, 1986;Chakraborty et al., 1998;Ueno et al., 2009;Polando et al., 2018). However, more recently, their roles have been contested based on several lines of evidence. First, MR ligands can also inhibit the uptake of L. donovani and L. major in the absence of MRs (Akilov et al., 2007). Secondly, inhibition of parasite uptake by MR ligands does not occur for all stages of promastigotes (Ueno et al., 2009) or all species (Da Silva et al., 1989;Mosser and Handman, 1992), and, finally, MR-depleted C57BL/6 macrophages do not differ in their infection rate compared to wild-type cells with regard to various aspects of the disease outcome, suggesting the existence of redundant infection receptors (Akilov et al., 2007). One explanation for these conflicting findings may be that mannose-carrying ligands in promastigotes can also bind to other C-type lectin receptors (McGreal et al., 2005) and Toll-like receptor (TLR) family members (Roeder et al., 2004). Another PRR, TLR-2, has also been implicated in the binding and internalization of L. major by BALB/c macrophages through its interaction with lipophosphoglycan (LPG) (Srivastava et al., 2013), a highly prevalent surface glycoconjugate in promastigotes.
Although, the direct binding of promastigotes to macrophages that occurs in vitro could also take place in vivo, the latter scenario is considerably more complex. Thus, it is very likely that opsonization may have a significant role in parasite uptake, as they most probably interact first with other host molecules, which would facilitate phagocytosis. In this regard, a major molecule is complement component 3 (C3), which, in vitro, mainly binds to gp63 and LPG after complement activation (Podinovskaia and Descoteaux, 2015;Sacks, 1992). The generated fragment C3b has been reported to be essential for promastigote entry into macrophages through binding to CR1 (Da Silva et al., 1989), while iC3b, a product of C3b cleaved by gp63, opsonizes the parasites by binding to CR3 (Brittingham et al., 1995). Other opsonizing molecules, including fibronectin (Vannier-Santos et al., 1991), C-reactive protein (Bodman-Smith et al., 2002) and heparan sulfate (Maciej-Hulme et al., 2018) have also been implicated in the internalization of promastigotes by macrophages.
Moreover, and importantly, even though macrophages are the final and main destination of parasites during the chronic phase of leishmaniasis, and may actually be directly infected by promastigotes in vivo (Peters et al., 2008), neutrophils are the major cells to capture promastigote during the initial phase of infection (Peters et al., 2008). In this case, macrophages may also become secondarily infected by ingesting Leishmania-containing apoptotic bodies after neutrophil death. This appears to be crucial for the survival of parasites and the outcome of infection (Afonso et al., 2008) and has been earlier referred to as the 'Trojan horse strategy' (Laskay et al., 2003).
Dendritic cells (DCs) are also targets for Leishmania spp. invasion. However, invasion of DCs is reported to occur mostly at later phases of the disease when susceptible promastigotes had either died through complement-mediated lysis or transformed into amastigotes inside macrophages, and when antibodies to parasites appear in sera. In fact, it has been shown that DCs preferentially take up amastigotes, rather than promastigotes, requiring opsonization through their receptor FcγR, which recognizes IgG (Guy and Belosevic, 1993;Peters et al., 1995;Kima et al., 2000).
Much less is known about the direct invasion of amastigotes or their transfer from one host cell to another, although this has been reported to occur via macrophage receptors for Fcγ or phosphatidylserine (Love et al., 1998;Kane and Mosser, 2000;De Freitas Balanco et al., 2001;Wanderley et al., 2006), phagocytosis of infected apoptotic bodies (Peters et al., 2008) or by direct cell-to-cell transfer of PV extrusions (Real et al., 2014).
Although the inhibition of phagocytosis mostly blocks invasion of macrophages by promastigotes, a small number of parasites can still be found inside these cells (Roy et al., 2014;Lewis, 1974;Aikawa et al., 1982), indicating that Leishmania can invade macrophages by routes other than classical phagocytosis. In fact, it has long been described that promastigotes can also invade nonphagocytic cells (Bogdan et al., 2000;Minero et al., 2004;Holbrook and Palczuk, 1975;Schwartzman and Pearson, 1985). Recent work from our group has unveiled a new non-phagocytic route of promastigote invasion without any involvement of the host cell cytoskeleton, thereby providing definitive evidence for an entry mechanism other than phagocytosis (Cavalcante-Costa et al., 2019). This pathway involves the Ca 2+ -dependent recruitment and exocytosis of host cell lysosomes to the parasite invasion site, where they instantly fuse with the PM, much like in the invasion mechanism described above for T. cruzi. This creates a niche for the internalization of promastigotes, which most often occurs through the flagella (Cavalcante-Costa et al., 2019) (Fig. 2C). It is possible, therefore, to speculate that, in vivo, Leishmania does not solely rely on being captured by phagocytes to establish infection, and that different infected cell types of the dermis, as already reported by other groups (Locksley et al., 1988;Bogdan et al., 2000), could act as Trojan horses.

Microsporidians
Microsporidians have long been considered as primitive protozoans with which they share some morphological similarities, but more accurate phylogenetic analysis have shown that they are, in fact, related to fungi (Weiss et al., 1999;James et al., 2006). These parasites were described by Louis Pasteur as a plague that decimated silkworms with a huge impact in silk industry in France, almost 150 years ago (Pasteur and Pasteur, 1870). Microsporidians are spread through parasite spores, which can infect not only several animals of economic importance, but also humans, being important pathogens. The phylum Microsporidia is composed by over 140 genera, but only eight have been characterized as causative of human microsporidiosis: Enterocytozoon, Encephalytozoon, Pleistophora, Trachipleistophora, Vittaforma, Brachiola, Septata and Nosema (Weiss, 2000). In humans, these parasites are able to infect almost every organ system causing asymptomatic, chronic or lethal infections, depending on the immunological status of the patient (Weber et al., 2000(Weber et al., , 1994 (Box 1).

Cell invasion by microsporidians
Microsporidian spores are rigid and round, and range from 1 to 12 µm in size, depending on the species; they can be divided in three basic morphological structures: the spore, the sporoplasm, which is the parasite body, and the invasion apparatus (Han and Weiss, 2017). The spore wall is mainly composed of chitin and glycoproteins (Han and Weiss, 2017). The most important family of proteins that form the spore wall are named spore wall proteins (SWPs), and these are essential for the formation of the spore but also participate in the adhesion of the spore to host cell PM before the formation of the invading apparatus, named the polar tube (recently reviewed by Yang et al., 2018) (Figs 3, 4D). The polar tube is a coiled structure, which is extruded in a fraction of a second, as a result of osmotic changes inside the spore (Fig. 3A,B). It can be as long as 500 µm, depending on the species, with a narrow diameter of between 0.1 and 0.2 µm, through which the sporoplasm travels to infect the host cell (Weidner, 1976;Han and Weiss, 2017). Polar tube proteins (PTPs), which form the polar tube, are the main factors involved in invasion. At least five PTPs have been described in microsporidians, with PTP1, PTP2, PTP3 and PTP5 found evenly distributed throughout the polar tube, whereas PTP4 is specifically located to the polar tube tip, in close interaction with host cell PM at the infection synapsis . PTP4 binds to transferrin receptor 1 (TfR1, also known as TFRC) on the PM of mammalian cells, and blocking this binding with specific antibodies has been shown to reduce infections in vitro, making this protein a promising new target for drug and vaccine development reviewed by Han and Weiss, 2017).
Even though microsporidians have been known about and studied for more than a century, essential features, such as the mechanism of direct injection into the cytoplasm and the endocytosis of the infective sporoplasm by the host cell still remain a matter of discussion. Interestingly, it has been shown that if the spore is phagocytosed by macrophages, the extrusion of the polar tube, typically termed germination, occurs from within the PV, leading to the perforation of the vacuolar membrane and the delivery of the infectious sporoplasm directly into the cytosol (Franzen, 2005); this constitutes a means to evade the deleterious conditions within fusogenic vacuoles. It is also interesting to note that microsporidians are able to replicate inside PVs, which are modified by the parasite to escape lysosomal acidification and to allow nutrient uptake from the cytoplasm (Rönnebäumer et al., 2008). Indeed, the membrane of the PV that contains Encephalitozoon cuniculi has been shown to originate from the host cell PM (Fig. 3C), as demonstrated by the addition of dyes to the host cell PM, which were found in the PV membrane shortly after cell invasion (Rönnebäumer et al., 2008). As PTP4 localizes to the tip of the polar tube and is able to interact with the host cell PM, Han et al. have proposed a model, in which PTP4 together with PTP1 helps forming the invagination of the host cell PM that excludes the extracellular environment, thus creating a microenvironment that is protected from host innate immune responses, thereby facilitating invasion . However, the machinery for pinching off and fission, which subsequently detaches the nascent membrane of the PV is still unknown. It has been hypothesized that invasion could also involve the clathrin mediated-endocytosis machinery , or host cell actin polymerization (Foucault and Drancourt, 2000). Although these obligatory intracellular parasites have the smallest genome among eukaryotes, they have evolved one of the most remarkable machinery to invade host cells. In this context, a better understanding of the proteins involved in spore adhesion to host cell PM (SWPs) and the ones involved in the functioning of the polar tube (PTPs), as well as the discovery of the molecular actors involved in PV pinch-off are crucial to unveil parasite entry process.

Concluding remarks
Even though the parasites discussed here have been known and studied for decades, some key issues of their biology and the details of the mechanism by which they invade host cells remain elusive. Many questions still remain open with regard to the invasion processes presented here. For instance, how exactly do nascent PVs pinch off from the PM and which membrane fission machinery is responsible for this crucial invasion step? Is membrane fission driven by the parasite itself, or does it require the recruitment of host cell molecules? Some attempts to solve this question may be still controversial. For instance, results drawn from inhibition experiments with dynasore might be ambiguous in inferring that dynamin participates in this process owing to the cytotoxic and offtarget effects of this drug (Preta et al., 2015). Therefore, more advanced approaches, such as the use of CRISPR/Cas9 or RNAi silencing to target candidates, such as dynamin itself, as well as other host cell proteins involved in membrane fission are anticipated to provide more reliable results. Importantly, as invasion does not occur without the initial adhesion of the parasite to the host cell, the discovery of the underlying molecules is also crucial. Another point that merits additional discussion is how infective Leishmania promastigotes that are delivered by sand flies actually reach macrophages. We know today, for instance, that after infection of mice with promastigotes, macrophages are not necessarily the first cells that are invaded by the parasite (Peters et al., 2008), and that the indirect infection of macrophages through phagocytosis of amastigotes, expelled from infected cells from PVs or as apoptotic bodies, or other mechanisms, are more likely to occur. Moreover, most conditions of in vitro infections, which are used to study the cell invasion process, do not recapitulate the in vivo situation, for instance, the use of fresh serum as a source of complement, because C3 opsonization of promastigotes, which is likely to occur in vivo, strongly increases their ability to infect macrophages. Hence, findings based solely on in vitro infection of macrophages by promastigotes should be interpreted with caution. Thus, more physiologically relevant approaches, such as in situ visualization of mice infections by intravital microscopy, using labeled cells and molecules, are needed to reach a better understanding of the invasion processes by these intracellular parasites. Similarly, experiments aiming to avoid off-target drug effects and to more closely reflect the processes occurring in nature would certainly be  Fig. 4. Overview over the cell invasion mechanisms employed by apicomplexans, kinetoplastides and microsporidians to entry host cells. (A) (1) Apicomplexans (e.g. T. gondii and Plasmodium spp.) first weakly interact with the host cell PM using their resident-GPI anchored proteins. At the same time, the strong interaction provided by microneme-secreted protein (MICs) leads to zoite re-orientation.
(2) With the apical complex facing the host cell PM, proteins secreted by the rhoptries interact with host cell proteins and (3) form the moving junction (see Fig. 1), which provides the force that drives parasites into the host cell cytoplasm. Inside host cells, parasites reside in a non-fusogenic vacuole and so avoid lysosomal fusion. (B) T. cruzi invades non-phagocytic host cells by inducing both (1) extracellular Ca 2+ influx through parasite-induced PM wounds and the release of intracellularly stored Ca 2+ .
(2) This leads to the recruitment of host cell lysosomes that function as Ca 2+ -sensitive exocytic vesicles, which donate membranes to the nascent PV.
(3) Inside host cells, the vacuole keeps fusing with host cell lysosomes, leading to (4) parasites escaping into the host cell cytosol where they replicate as round-shaped amastigotes. (C) Leishmania spp. are captured by phagocytes through either (1) actin-mediated phagocytosis or (2) by non-phagocytic cells through Ca 2+ -dependent recruitment of host cell lysosomes that donate membranes to the nascent PV. This process appears to be the same as infection by T. cruzi trypromastigotes, indicating that the underlying mechanisms are conserved in these kinetoplastides. After invasion, Leishmania spp. live and replicate inside acidic vacuoles that fuse with host cell lysosomes. (D) Microsporidian spores bind to host cells with their spore wall and extrude their polar tube, which creates an invagination in the host cell PM. The infective sporoplasm that harbors the spore travels along the polar tube lumen and is delivered at the infection site. After internalization, the parasite lives and replicates inside nonfusogenic vacuoles and so avoids lysosomal fusion.
helpful in making appropriate choices for molecular targets used in drug discovery and vaccine development. Finally, some features of the invasion processes are common among phylogenetically-related parasites. For example, cell invasion by Plasmodium spp. and Toxoplasma gondii zoites may differ in terms of the specific host cell invaded, initial adhesion molecules or proteins used for the assembly of the MJ, but, ultimately, the same invading machinery is formed. Similarly, we now know that the infective stages of both T. cruzi and Leishmania are able to induce cell invasion by subverting PMR-induced endocytosis (Fig. 4). Thus, building on established knowledge, extrapolation of models and mechanisms from one species could help to narrow knowledge gaps in closely related microorganisms, thereby accelerating new discoveries and guiding the proposal of new approaches to better understand the biology of these parasites and the diseases they cause.