The Molecular Basis of Erythrocyte Invasion by Malaria Parasites

Abstract

Plasmodium species cause malaria by proliferating in human erythrocytes. Invasion of immunologically privileged erythrocytes provides a relatively protective niche as well as access to a rich source of nutrients. Plasmodium spp. target erythrocytes of different ages, but share a common mechanism of invasion. Specific engagement of erythrocyte receptors defines target cell tropism, activating downstream events and resulting in the physical penetration of the erythrocyte, powered by the parasite’s actinomyosin-based motor. Here we review the latest in our understanding of the molecular composition of this highly complex and fascinating biological process.

Introduction

Despite increased control measures, malaria remains a major public health burden. Currently more than 3 billion people are at risk, with an estimated 200 million infections and more than 400,000 deaths each year (WHO, 2016). Plasmodium spp., the causative agent of malaria, belongs to a larger phylum of obligate intracellular parasites, the Apicomplexa. There are six species that cause malaria in humans: Plasmodium falciparum, P. vivax, P. knowlesi, P. ovale curtisii, P. ovale wallikeri, and P. malariae. P. falciparum is considered the most important because it causes the vast majority of deaths; however, more recently P. vivax mortality and morbidity has been reassessed and are likely to have been significantly underestimated (Naing et al., 2014). Malaria due to P. ovale curtisii, P. ovale wallikeri, and P. malariae are much less common while P. knowlesi has become an important cause of malaria in Southeast Asia. P. knowlesi is predominantly a zoonotic parasite of Southeast Asian macaques, with as yet no definitive evidence of primary human-to-human transmission (Ahmed and Cox-Singh, 2015).

The Plasmodium spp. life cycle involves two different hosts: the definitive host or vector, which is the mosquito of the genus Anopheles, followed by a vertebrate intermediate host, such as a human (Figure 1). Haploid sporozoites present in salivary glands of female mosquitoes are injected into the vertebrate host during a blood meal. The sporozoites rapidly migrate to the liver by traversing through cells and only actively invading hepatocytes. The intracellular parasite divides over a number of days to form many thousands of merozoite forms that are released into the bloodstream to invade erythrocytes and begin the asexual blood-stage cycle. Once inside the erythrocyte, the parasite remodels the host cell to provide a niche in which it can obtain nutrients required for development and division as well as provide a means to evade the host immune system (reviewed in Boddey and Cowman, 2013). Intracellular forms develop through the trophozoite and schizont stage to form 16–32 merozoites that are released during egress and are primed for invasion of new erythrocytes to continue the asexual blood cycle, amplifying exponentially and causing the disease symptoms of malaria (reviewed in Cowman et al., 2016).

Figure 1. The Life Cycle of P. falciparum in the Human Host and the Anopheles Mosquito Vector.

The Plasmodium-infected mosquito injects sporozoites into the host and these migrate to the liver, where they pass through Kupffer cells and invade hepatocytes within which they develop to liver merozoites. Liver merozoites are released into the bloodstream where they invade erythrocytes. They develop through ring, trophozoite, and schizont stages replicating to produce from 16–32 merozoites that are released at egress. The free merozoite invades new erythrocytes to continue the asexual blood-stage life cycle. Some intraerythrocytic stages develop into male or female microgametocyte and macrogametocytes, the sexual forms of the parasite. These are taken up by the mosquito during feeding and develop into gametes in the insect gut and fuse to form zygotes. The zygote develops to form an invasive ookinete, which traverses the midgut and transforms into an oocyst, from which sporozoites are released that migrate to the salivary glands for injection into a human host during the next blood meal.

Plasmodium spp. merozoites and sporozoites are uniquely adapted to respectively invade erythrocytes and hepatocytes, cell types to which they are uniquely restricted. This is in contrast to other Apicomplexa, such as Toxoplasma gondii with the invasive tachyzoite able to invade most cell types (reviewed in Cowper et al., 2012). In this review, we concentrate on the molecular mechanisms involved in invasion of erythrocytes by Plasmodium spp. merozoites. While we emphasize P. falciparum, about which most is known, we also discuss other Plasmodium spp., particularly with respect to host tropism. Additionally we discuss T. gondii because its genetic tractability and conserved molecular machinery has provided important aspects for understanding erythrocyte invasion by Plasmodium spp.

Mechanical Steps of Erythrocyte Invasion

The blood-stage merozoite of Plasmodium spp. is relatively small with a length and breadth of approximately 1–2 μm. Merozoites have a single goal—to invade the next erythrocyte—and as such are designed solely for this purpose. A polarized cell, the merozoite’s apical end contains organelles and structures, including micronemes and rhoptries, which facilitate invasion upon erythrocyte contact (Figure 2A). Micronemes contain adhesins involved in erythrocyte binding, while rhoptries are released after initial host cell engagement to facilitate the invasion process and form the parasitophorous vacuole where the merozoites replicate and form daughter cells. Recent evidence suggests that micronemes may come in several flavors depending on the adhesins they store, allowing for a highly organized program of release, while the club-shaped rhoptries are partitioned into subdomains with the parasite, perhaps allowing temporal release of factors (Healer et al., 2002; Kremer et al., 2013). Additionally, the dense granule organelles may also have distinct subpopulations, and there is a subset of organelles named exonemes that release the protease subtilisin 1 (SUB1) (Table 1) into the parasitophorous vacuole, where it proteolytically processes several parasite proteins to facilitate merozoite egress (Yeoh et al., 2007). Dense granules are released by fusion with the plasma membrane of the parasite at different stages in the asexual cycle, including during or just after invasion. The compartmentalization and subsets of these organelles are likely critical in enabling temporal release of specific ligands for each step of invasion.

Figure 2. Structure of the Merozoite and Steps in its Interaction with the Host Erythrocyte.

(A) The subcellular structure of a P. falciparum merozoite showing the microneme and rhoptry organelles at the apical end.

(B) Merozoite invasion of erythrocytes. Invasion involves an initial attachment, which may involve MSPs or directly with EBAs and PfRh proteins. Apical reorientation likely involves membrane wrapping so that the apical end is adjacent to the erythrocyte, allowing tight attachment. Evidence suggests that a pore is formed between the merozoite and erythrocyte that is mediated either directly or indirectly by the PfRh5/PfRipr/CyRPA complex. This is associated with movement of the RON complex on the host membrane. A tight junction is formed involving high-affinity ligand-receptor interactions between AMA1 on the merozoite surface and RON2 inserted in the erythrocyte membrane. This tight junction then moves from the apical to posterior pole powered by the parasite’s actinomyosin motor. The surface coat is shed at the moving junction by a serine protease, or “sheddase.” Upon reaching the posterior pole, the adhesive proteins at the junction are also proteolytically removed by a resident protease, most likely a rhomboid, in a process that facilitates resealing of the membrane. By this process the parasite does not actually penetrate the membrane, but invades in a manner that creates a parasitophorous vacuole.

Table 1.

Some Proteins Associated with and Involved in Invasion of Plasmodium Merozoites

Name	Accession No.	Function/Features	Structure
GPI-Anchored MSPa
MSP-1	PF3D7_0930300	required for merozoite egress through binding spectrin, major surface antigen and may bind Band 3; forms major complex on merozoite surface	two epidermal growth factor (EGF) domains in side-by-side arrangement at C terminus
MSP-2	PF3D7_0206800	not known
MSP-4	PF3D7_0207000	not known
MSP-5	PF3D7_0206900	not known
MSP-8	PF3D7_0502400	not known
MSP-10	PF3D7_0620400	not known	two EGF domains in side-by-side arrangement at C terminus
P12	PF3D7_0612700	member of 6-cys family, binds to Pf41	6-cys domains with similarity to T. gondii surface protein SAG1; P12 structure reveals two juxtaposed 6-cys domains
P38	PF3D7_0508000	member of 6-cys family	6-cys domains with similarity to T. gondii protein SAG1
P41		member of 6-cys family, binds to P12	P41 requires an interdomain insertion for interaction with P12
P92	PF3D7_1364100	cys-rich surface protein; binds factor H and involved in complement evasion
P113	PF3D7_1420700	member of Plasmodium translocon for exported proteins (PTEX); also reported to bind the amino terminus of PfRh5
Micronemes
AMA-1	PF3D7_1133400	binds to RON2 in the RON complex at the tight junction, implicated in signaling to activate subsequent steps	two closely associated PAN domains
EBA-175	PF3D7_1133400	binds glycophorin A, implicated in signaling release of rhoptries; implicated in signaling to erythrocyte making host membrane more deformable	two DBL domains with “handshake” association for glycophorin A glycan binding; paralog of PvDBP
EBA-181/JESBL	PF3D7_0102500	binds trypsin-resistant erythrocyte receptor W	two DBL domains; paralog of EBA-175
EBA-140/BAEBL	PF3D7_1301600	binds glycophorin C; implicated in signaling release of rhoptries; implicated in signaling to erythrocyte making host membrane more deformable	two DBL domains; paralog of EBA-175
EBL1	PF3D7_1371600	binds glycophorin B, pseudogene in most P. falciparum parasites	two DBL domains; paralog of EBA-175
PvDBP	PVX_110810	binds Duffy antigen/chemokine receptor (DARC) and is a host tropism determinant for P. vivax	single DBL domain
PkDBP	PKNH_0623500	binds Duffy antigen/chemokine receptor (DARC) and is a host tropism determinant for P. knowlesi	single DBL domain
ASP	PF3D7_0405900	no known function; contains sushi domain, hence name apical sushi protein; putative GPI anchor
PfRipr	PF3D7_0323400	forms complex with PfRh5 and CyRPA at interface between merozoite and erythrocyte during invasion; associated with formation of pore with host cell	cysteine-rich with 10 EGF-like domains
CyRPA	PF3D7_0423800	forms complex with PfRh5 and CyRPA at interface between merozoite and erythrocyte during invasion; associated with formation of pore with host cell	6-propeller structure
CLAMP	PF3D7_1030200	essential for asexual growth, likely involved in tight junction formation	claudin-like protein
Peripheralb
MSP9	PF3D7_1228600	no known function but putative protease
S-antigen	PF3D7_1035200	no known function, mainly released into the media at parasite egress
GLURP	PF3D7_1035300	no known function
MSP3	PF3D7_1035400	binds to the MSP-1 complex
MSP6	PF3D7_1035500	binds to the MSP-1 complex
MSP7	PF3D7_1335100	binds to the MSP-1 complex
MSP7-like protein family	PF3D7_1334400 PF3D7_1334500 PF3D7_1334700 PF3D7_1334800 PF3D7_1335000	some members may bind to the MSP-1 complex; unknown function
MSPDBL1	PF3D7_1035700	binds unknown receptor on erythrocyte through DBL domain; binds to the MSP-1 complex	structure of the DBL domain modeled on MSPDBL2
MSPDBL2	PF3D7_1036300	binds unknown receptor on erythrocyte through DBL domain; binds to the MSP-1 complex	structure of the DBL domain solved
H101	PF3D7_1035600	binds to the MSP-1 complex; MSP3-like protein
H103	PF3D7_1036000	Binds to the MSP-1 complex; MSP3-like protein
SERA3	PF3D7_0207800	cysteine protease domain with active site serine; unknown function
SERA4	PF3D7_0207700	cysteine protease domain with active site serine; unknown function
SERA5	PF3D7_0207600	cysteine protease domain with active site serine; unknown function
SERA6	PF3D7_0207500	cysteine protease domain with active site serine; unknown function
Rhoptry Neck
PfRh1	PF3D7_0402300	binds unknown receptor Y on erythrocyte, implicated in signaling for release of rhoptries; implicated in signaling to erythrocyte to activate a phosphorylation cascade making host membrane more deformable
PfRh2a	PF3D7_1335400	no demonstrated function
PfRh2b	PF3D7_1335300	binds unknown receptor Z on erythrocyte; implicated in signaling for release of rhoptries; implicated in signaling to erythrocyte to activate a phosphorylation cascade making host membrane more deformable
PfRh3	PF3D7_1252400	probable transcribed pseudogene
PfRh4	PF3D7_0424200	binds complement receptor 1 (CR1, CD35) on erythrocyte; implicated in signaling for release of rhoptries; implicated in signaling to erythrocyte to activate a phosphorylation cascade making host membrane more deformable
PfRh5	PF3D7_0424100	binds basigin (CD147) on erythrocyte surface; binds CyRPA to form tripartite complex with PfRipr at interface of merozoite and erythrocyte during invasion; binds P113 to tether to the membrane, associated with formation of pore with host cell; host tropism determinant	novel structure consisting predominantly of α helices
PvRBP	PVP01_1402400 PVX_121920	PvRBP1a and 2c form a high molecular weight complex that binds reticulocytes; PvRBP2a binds normocytes, PvRBP2b binds reticulocytes	structure of the PvRBP2a has a similar fold to PfRh5
RON2	PF3D7_1452000	injected into erythrocyte membrane during invasion; binds to AMA-1 at the tight junction; forms a complex with AMA-1, RON2, RON4, and RON5	two membrane spanning domains; region protruding from the erythrocyte membrane inserts into a groove of AMA-1
RON4	PF3D7_1116000	injected under erythrocyte membrane during invasion; forms complex with RON2 and RON5
RON5	PF3D7_0817700	injected under erythrocyte membrane during invasion; forms complex with RON2 and RON4
DGc
SUB1	PF3D7_0507500	subtilisin-like serine protease
SUB2	PF3D7_1136900	subtilisin-like serine protease; MSP-1/AMA-1 processing “sheddase”

^aMerozoite surface proteins (MSPs) known or predicted GPI anchors (see Sanders et al., 2005).

^bPeripheral proteins associated with the merozoite surface.

^cDG, dense granules.

While the apical organelles play an essential role in invasion, it is the surface of the merozoites that first contacts the erythrocyte, activating a series of mechanical steps that end in active entry into the host cell. These steps are rapid and were first visualized using video-microscopy of P. knowlesi merozoites (Dvorak et al., 1975) and subsequently defined for P. falciparum (Gilson and Crabb, 2009) to show that the sequence and kinetics of this process are conserved across Plasmodium spp. (Figure 2B). This has divided invasion into three phases: first, initial interaction and deformation of the erythrocyte membrane; second, apical interaction and invasion; and third, echinocytosis (rapid shrinkage of the erythrocyte with evenly spaced projections) and recovery of the invaded host cell (Figure 2B). Previously, it has been unknown whether the erythrocyte actively participates in the invasion process; however, modeling of the merozoite interaction has suggested that wrapping of the erythrocyte cell membrane followed by reorganization of the underlying cytoskeleton could account for all the energetic steps required for the invasion process (Dasgupta et al., 2014). Furthermore, it has been shown that merozoite contact results in increased phosphorylation of the erythrocyte cytoskeleton (Zuccala and Baum, 2011). Additionally, binding of ligands to the erythrocyte surface causes changes in membrane deformability by activating a phosphorylation cascade, involving the kinase activity of the host transient receptor potential cation channel TRPM7 (Sisquella et al., 2017). Moreover, real-time deformability cytometry and flicker spectroscopy has shown that binding of the parasite ligand EBA175 to its host receptor, glycophorin A (GPA), increases erythrocyte cytoskeletal tension and reduces the bending modulus of the cell’s membrane, which correlate with efficiency of merozoite invasion (Koch et al., 2017). These studies are consistent with the interaction of merozoite ligands (Table 1) and receptors on the erythrocyte activating a phosphorylation cascade that involves the host cytoskeleton and alters the viscoelastic properties of the membrane, a process important for invasion.

Molecular Basis of Erythrocyte Invasion

Merozoite surface proteins (MSPs) could play a role in the first phase of invasion involving the initial interaction of the merozoite with the erythrocyte (Figure 2B). MSP1 forms a large complex on the merozoite surface with a number of peripheral proteins (Table 1) and there is evidence that it is required for invasion. However, recent data have shown that merozoites lacking MSP1 expression can still invade, suggesting it may not be essential for this process (Das et al., 2015). Indeed, MSP1 appears to be required for egress from the erythrocyte in which the SUB1 processed form of MSP1 interacts with the spectrin cytoskeleton of the erythrocyte. As a result it is unclear which, if any, of the MSPs play a role in the initial phase of invasion.

While surface and peripheral proteins may not be involved directly in the invasion process, some do play a role in protecting the merozoite from immune attack, and they are also targets of naturally acquired immunity. Once released from the infected erythrocyte, merozoites are exposed to complement attack. A recent study in P. falciparum has shown Pf92, a member of the 6-cys family, actively recruits Factor H (FH), a major negative regulator of the alternative pathway of complement activation (Kennedy et al., 2016). Hijacking FH allows the merozoite to downregulate complement activation on its surface, consequently protecting it from complement lysis. On the other hand, MSPs are also targets of naturally acquired immunity, providing a repertoire of human antibodies that prevent parasite invasion. Human antibodies to MSP1 and MSP2 promote complement deposition on the merozoite and mediate inhibition of parasite invasion through C1q fixation (Boyle et al., 2015). This antibody-mediated complement-dependent inhibition appears to be an important mechanism by which most invasion-inhibitory human antibodies function in preventing malaria parasites from entering erythrocytes. Collectively, these studies highlight the complex interplay between the merozoite and human complement system and emphasize the need to study parasite invasion in the presence of complement to understand its immune evasion mechanisms as well as how invasion-inhibitory antibodies function beyond blocking ligand-receptor interactions (Schmidt et al., 2015).

It is clear that initial tight interaction with the erythrocyte is mediated by two major families of adhesins released most likely from the micronemes upon intracellular signaling (see below). These are the Duffy binding-like (DBL or erythrocyte-binding-like [EBL]) protein and the reticulocyte-binding-like protein homolog (Rh or RBL) (Figure 3A, Table 1). The first DBL protein was identified in P. knowlesi and shown to bind to Duffy antigen/chemokine receptor (DARC) (Haynes et al., 1988), and orthologs in P. vivax and P. falciparum have been identified. In P. falciparum, this family consists of EBA-175 (binds to GPA) (Camus and Hadley, 1985), EBA-181 (also known as JSEBL) (unknown receptor) (Gilberger et al., 2003), EBA-140 (also known as BAEBL) (binds to glycophorin C [GPC]) (Maier et al., 2003; Mayer et al., 2006) and EBL1 (binds to glycophorin B [GPB]) (Mayer et al., 2009) (Table 1 and Figure 3A). The RBL or Rh family was first identified in P. vivax (Galinski et al., 1992) and is now known to have orthologs in other Plasmodium spp. (Galinski et al., 1992). In P. falciparum this includes PfRh1 (binds to unknown receptor), PfRh2a (binds to unknown receptor), PfRh2b (binds to unknown receptor) (Duraisingh et al., 2003; Rayner et al., 2001; Stubbs et al., 2005; Triglia et al., 2005), PfRh4 (binds to complement receptor 1 [CR1, CD35]) (Tham et al., 2010), and PfRh5 (binds to basigin [BSG, CD147]) (Crosnier et al., 2011) (Figure 3A). While most of these proteins individually appear not to be essential, the function of them overall is required for invasion (Lopaticki et al., 2011). The exception is PfRh5, a protein that is much smaller than other PfRh proteins without a transmembrane region and responsible for a separate and essential step in invasion (Volz et al., 2016; Weiss et al., 2015).

Figure 3. Parasite Ligand-Receptor Interactions and Events Associated with the Actomyosin Motor.

(A) Erythrocyte receptor and parasite ligand interactions involved in merozoite invasion for P. falciparum, P. vivax, and P. knowlesi.

(B) A model for the merozoite motor and associated proteins based mainly on evidence from T. gondii tachyzoites. The adhesin complex is linked to actin through the glideosome associated protein (GAC). The MyoA neck region is stabilized through the essential light chain (ELC) and linked to the glideosome associated protein 45 (GAP45). GAP45 spans the inner membrane complex (IMC) and the plasma membrane and is associated with GAP40 and GAP50 in the IMC. Short actin filaments are depolymerized by cofilin to G-actin. Activity of the glideosome is regulated at the apical end of the zoite where the actin nucleating protein formin1 is localized. GAC is held in check by the apical lysine methyl transferase (AKMT) at the apical end. G-actin is polymerized into F-actin via profilin. Adhesins are released from micronemes and phosphatidic acid produced during signaling events.

Many ligand-receptor interactions remain to be characterized, as many of the receptors for EBAs/PfRh have not been identified (Figure 3A). In particular, a comprehensive small hairpin RNA knockdown screen identified decay accelerating factor (DAF, CD55) as an erythrocyte receptor required for an essential step in parasite invasion (Egan et al., 2015). Identification of its parasite ligand will allow a novel focus on development of new therapeutics that would halt parasite invasion. As DAF plays a role in complement regulation on the erythrocyte surface, this highlights a need to study the process of parasite invasion in an environment of active complement.

Live imaging of merozoites in the presence of specific inhibitors has shed light on the distinct steps of invasion executed by different receptor-ligand interactions including the DBL and Rh ligands (Weiss et al., 2015). When the merozoite initiates contact with the erythrocyte it weakly buckles, leading to stronger deformation of the erythrocyte through engagement of the EBAs/PfRhs with their receptors, such as CR1, GPA, GPB, and GPC (Figure 3A and Table 1), and the involvement of the actinomyosin motor (Figure 3B). Merozoites that strongly deform the erythrocyte surface are more likely to successfully invade, leading to the suggestion that deformation allows the merozoite to embed itself securely into the erythrocyte surface and resist detachment under blood-flow conditions. During this time, merozoites also use parasite calcineurin, a protein phosphatase complex that responds to Ca²⁺ signaling that initiates during merozoite-erythrocyte contact, to strengthen the host-pathogen adhesions mediated by EBAs/PfRhs and the erythrocyte receptors (Paul et al., 2015) (Figure 3B). Intriguingly, calcineurin has not been implicated in the steps of initial attachment but to function together with EBAs/PfRh and apical membrane antigen 1 (AMA1), suggesting that this signal transduction pathway is linked to host cell recognition (Paul et al., 2015).

Strong deformation precedes apical reorientation of the merozoite, most likely induced by erythrocyte membrane wrapping (Dasgupta et al., 2014) (Figure 2B) and PfRh5 engagement with its host receptor basigin (Crosnier et al., 2011) (Figure 3A). This essential step in invasion is linked to a Ca²⁺ flux that has been suggested to emanate from the merozoite into the erythrocytes, through a pore that may be formed during PfRh5 engagement (Volz et al., 2016; Weiss et al., 2015) (Figure 2B). While definitive proof that the Ca²⁺ flux emanates from the merozoite, rather than the medium, is still lacking, it does support the idea that a pore is formed between the invading merozoite and the erythrocyte membrane through which Ca²⁺ and, perhaps, parasite proteins can be inserted into the host cell (Figure 2B). This Ca²⁺ flux is strongly correlated with the observation of echinocytosis, the further release of the merozoite rhoptry contents, and successful invasion (Weiss et al., 2015). During invasion, PfRh5 forms a complex with PfRipr (P. falciparum Rh5 interacting protein) and CyRPA (cysteine-rich protective antigen) (Volz et al., 2016) (Figure 3A), and knockdown analyses of these show that both are also required for the generation of the Ca²⁺ flux and play an essential role in parasite invasion (Chen et al., 2011; Reddy et al., 2015; Volz et al., 2016). While PfRh5, PfRipr, and CyRPA have different spatial localizations in schizonts and merozoites, super-resolution microscopy shows consistent co-localization of all three proteins at the merozoite apical tip when the parasite membrane is in close juxtaposition with the erythrocyte membrane (Chen et al., 2011; Reddy et al., 2015; Volz et al., 2016). This suggests that co-localization of the tripartite PfRh5-PfRipr-CyRPA complex at the junction between the merozoite and erythrocyte is either directly or indirectly required for formation of a pore and a signal for subsequent steps in invasion such as tight junction formation.

During parasite invasion, PfRh5 and PfRipr are tightly associated with a membrane, suggesting they must be tethered to the merozoite surface; however, these proteins lack a transmembrane domain or GPI anchor. A recent study suggested that CyRPA was GPI anchored (Reddy et al., 2015), but this has not been verified (Galaway et al., 2017; Volz et al., 2016). Pf113, an abundant GPI-anchored protein suggested to be present on the merozoite surface, was identified as the membrane tether for PfRh5 (Galaway et al., 2017) (Figure 3A). It has been shown that Pf113 interacts with the amino terminus of PfRh5 and forms a complex together with basigin (Galaway et al., 2017). However, there is no evidence that Pf113 interacts with the ternary PfRh5-PfRipr-CyRPA complex and no functional evidence that it is involved in merozoite invasion. Additionally, Pf113 appears to be located in dense granules associated with PTEX, a translocon that exports proteins from the parasitophorous vacuole into the host cytosol (Elsworth et al., 2016), and it seems unlikely it is present at the junction between the invading merozoite and erythrocyte. Nevertheless, if Pf113 does anchor PfRh5 to the merozoite surface, suggesting a molecular model for it to interact with the receptor basigin. The subsequent binding of CyRPA and PfRipr to PfRh5 would release PfRh5 engagement with Pf113, allowing the PfRh5-PfRipr-CyRPA ternary complex to execute the next step in invasion, which may involve formation of a pore between parasite and erythrocyte membranes (Volz et al., 2016; Weiss et al., 2015).

Tight junction formation occurs downstream of PfRh5 engagement (Figure 2B). During this stage in invasion, the merozoite moves through the tight junction propelled by its actinomyosin motor to enter the erythrocyte. At its core the tight junction is composed of AMA1 and RON2 (Alexander et al., 2005; Besteiro et al., 2009; Riglar et al., 2011; Tonkin et al., 2011) (Figure 3A). AMA1 is found on the surface of the merozoite and binds to RON2, which is part of a larger RON complex, and is one of the first proteins to be injected into the erythrocyte, emanating from the rhoptry neck. Work from the related parasite T. gondii suggests that RON2 embeds itself into the host membrane, becoming the receptor for AMA1, an ingenious mechanism used by all apicomplexan parasites to deploy their own ligand-receptor pair to facilitate successful invasion. While there has been some controversy over the essentiality of AMA1 to parasite invasion, the vast body of evidence suggests that its interaction with the RON complex is highly important (Bargieri et al., 2013). Interestingly, if tight junction formation is inhibited by blocking the AMA1-RON2 interaction, erythrocyte echinocytosis is still observed, strongly suggesting that tight junction formation lies downstream of PfRh5 engagement and the signaling events promoted by rhoptry release (Weiss et al., 2015). Recent data using an elegant genome-wide CRISPR screen in T. gondii has identified claudin-like apicomplexan microneme protein (CLAMP), a protein that has similarity to mammalian tight junction proteins and may serve in this role (Sidik et al., 2016). Nonetheless the role of this protein in P. falciparum is unclear, although loss-of-function mutants are unable to invade and do not form a tight junction, suggesting that it also plays a key role in this process.

Characterization of the multiple ligand-receptor interactions involved in invasion has led to the development of antibodies and inhibitors that effectively block parasite invasion. In particular, antibodies that block function of PfRh5 and its partners show the most efficacious inhibition in both in vitro and in vivo models of parasite invasion (Chen et al., 2011, 2017; Douglas et al., 2015, 2011, 2014; Dreyer et al., 2012; Favuzza et al., 2017). While EBAs/PfRhs function in a distinct step in parasite invasion from PfRh5, antibodies to EBA-175 and PfRh4 together with PfRh5 antibodies show an additional synergistic inhibition in parasite invasion, suggesting that further understanding of the interdependencies of these events are needed (Williams et al., 2012). Collectively, the invasion-inhibitory studies show that targeting multiple steps in invasion is more likely to promote higher levels of parasite invasion inhibition, suggesting that a multivalent vaccine will be more effective than a single subunit vaccine against P. falciparum (Lopaticki et al., 2011; Williams et al., 2012).

Signal Transduction within the Merozoite

Invasion is a tightly regulated process relying on rapid signal transduction, and progress has been made in identifying how parasite adhesins are deployed in invasion. Pharmacological inhibition studies implicate both Ca²⁺ and phosphoinositide signaling as playing key roles in Plasmodium spp. invasion (Dawn et al., 2014). Proteomics studies indicate extensive changes in the phosphoproteome when merozoites are primed for invasion (Lasonder et al., 2012).

Following egress, external signals result in Ca²⁺ release from internal stores and in microneme exocytosis. It is thought, although it remains unproven, that binding of EBAs on the merozoite surface to erythrocyte receptors results in rhoptry release (Singh et al., 2010). In this mechanism, sensing of the low K⁺ environment leads to an increase in Ca²⁺ levels within the merozoite. Using small-molecule inhibitors, a pathway has been postulated (Dawn et al., 2014) in which bicarbonate levels are increased by the action of the carbonic anhydrase PfCA, which activates the adenylyl cyclase PfACbeta to produce cyclic AMP (cAMP). cAMP in turn activates protein kinase A (PfPKA) that mediates some processes of invasion, as well as P. falciparum exchange protein activated by cAMP (PfEPAC) (Dawn et al., 2014), which activates phospholipase C (PLC) to release Ca²⁺. However, a huge gap exists in our knowledge of the mechanisms of Ca²⁺ release, with IP₃ and ryanodine receptors not being obviously encoded within the P. falciparum genome.

Other studies suggest that it is engagement of EBA and PfRhs with their receptors that activate signaling pathways of invasion, and this interaction precedes microneme secretion (Gao et al., 2013; Tham et al., 2015). In this model, the first signaling event within merozoites occurs following engagement of EBA and Rh proteins with receptors on the surface of the erythrocyte. Most members of the EBA and Rh families are type I transmembrane proteins, and it is postulated that engagement results in dimerization and transmission of a signal. Dimerization has been demonstrated at the structural level for the EBA-175/GPA interaction (Tolia et al., 2005), while signaling resulting from this dimerization has yet to be demonstrated. Interestingly, the EBA140 ligand binds in a monovalent fashion to its receptor GPC (Tolia et al., 2005), which is lowest in the hierarchy of invasion ligand-receptor interactions (Baum et al., 2005). Phosphorylation of the tails of certain EBAs and Rhs has been observed (Engelberg et al., 2013; Tham et al., 2015). Evidence has been presented that this phosphorylation requires the parasite kinase PfCK2, and is required at least for PfRh4 function (Tham et al., 2015).

Signaling by the AMA1 protein is thought to be downstream of this primary signaling event, and phosphorylation of its tail through PfPKA is clearly required for function (Leykauf et al., 2010) and is involved in sequential phosphorylation steps (Prinz et al., 2016). Indeed, phosphoproteomics analysis has clearly implicated PKA and inositol phosphate signaling as major regulators in egress and invasion (Lasonder et al., 2012).

While it is not known how phosphorylation of ligand-receptor interactions transmit signals to elicit the next steps in invasion, genetic approaches identified essential members of the signal transduction machinery responsive to Ca²⁺. A plant-like kinase family, the calcium-dependent protein kinases (CDPKs), has been identified in apicomplexans, including Plasmodium spp., which possess a kinase domain fused to Ca²⁺-binding domains. While PfCDPK5 plays a major role in egress from erythrocytes (Dvorin et al., 2010), evidence suggests that PfCDPK1 is required for invasion (Holder et al., 2012) as well as egress (Kato et al., 2008). PfCDPK1 may regulate microneme secretion, and also the actinomyosin motor, through phosphorylation of glideosome-associated protein 45 (PfGAP45; discussed further below), which plays essential roles in parasite motility during entry (Green et al., 2008). A key Ca²⁺-regulated protein mediating microneme release is double C2 domain protein PfDoc2, which is essential for invasion (Farrell et al., 2012). Similarly, a role for Plasmodium protein kinase G (PfPKG) in both egress and invasion has been demonstrated through specific inhibition of PfPKG (Taylor et al., 2010), with phosphoproteomics revealing signaling hubs around these regulators (Alam et al., 2015). While we have reviewed the effectors of signal transduction in P. falciparum merozoite invasion, this is likely to be related to the mechanisms used for the parasite egress from erythrocytes that directly precedes invasion (Lew and Tiffert, 2007).

Powering Invasion

After receptor-adhesin engagement Plasmodium spp., like all extracellular apicomplexan parasites, must propel themselves into the host cell. The current model posits that the “glideosome” links transmembrane adhesins (such as EBAs, RBPs/PfRhs, and AMA1) directly or indirectly to the inner membrane complex (IMC) and provides forward thrust by the actomyosin motor (Figure 3B). This proposes that adhesins are dragged through the plane of the membrane to the rear of the parasite, thus causing forward movement and invasion. Most of the insight into how invasion is regulated and powered has come from the work in the related and more experimentally tractable parasite, T. gondii, to which discoveries have proved almost universally translatable to Plasmodium spp. invasion and motility throughout the whole life cycle (Baum et al., 2006b; Lourido et al., 2012).

The glideosome is made up of the “glideosome-associated proteins” (GAPs), which act as a scaffold, providing an anchor point for myosin A (MyoA) (Meissner et al., 2002), while also keeping the parasite IMC that provides structural stability and the plasma membrane in close apposition (Frenal et al., 2010; Gaskins et al., 2004) (Figure 3B). GAP45, a putative, extended coiled-coil protein, spans the supra-aleveolar space anchoring, by myristoylation and palmitoylation, to both the IMC and the other end to the plasma membrane (Frenal et al., 2010). The related GAP40, GAP50, and GAPM proteins are embedded in the IMC and, while they appear to be important for IMC biogenesis, their exact role in glideosome function is not known (Harding et al., 2016). MyoA has two light chains, MTIP (known as MLC1 in T. gondii) and the redundant essential light chain 1 (ELC1) and ELC2 (the orthologs are yet to be identified in Plasmodium), which stabilize and rigidify the neck region (Bookwalter et al., 2014; Green et al., 2006; Williams et al., 2015). MyoA-MTIP and ELC complex then link to the GAP proteins using a unique extension on MTIP/MLC1 (Frenal et al., 2010) (Figure 3B).

Critical to invasion and motility is the regulation of actin dynamics. Apicomplexan parasites have a greatly reduced repertoire of actin-regulating proteins, in particular an almost complete absence of proteins involved in actin branching and bundling, suggesting less complex mechanisms of regulation (Baum et al., 2006a). Indeed, it has been postulated that apicomplexan parasites produce short unstable filaments (Vahokoski et al., 2014). Yet studies have been hampered by concerns with respect to the correct folded state of the recombinantly expressed parasite actin and the inability to visualize native filaments in vivo (Olshina et al., 2016). What is clear is that actin polymerization factors such as the nucleator formin proteins and monomeric actin-binding protein profilin are crucial for motility and invasion (Baum et al., 2008; Daher et al., 2010; Plattner et al., 2008). Furthermore, cofilin, which is required for the depolymerization of actin filaments, is also required for parasite motility (Mehta and Sibley, 2010, 2011) (Figure 3B).

To produce the force required for invasion, the glideosome needs to connect to the cytoplasmic tails of adhesins. For a decade, evidence suggested that the glycolytic enzyme aldolase moonlighted to act as a connector between actin filaments and the cytoplasmic tail of adhesins (Jewett and Sibley, 2003). However, convincing new experimental data in T. gondii provides a more plausible mechanism and some new insight into regulation of motility (Jacot et al., 2016). An armadillo-repeat protein has been identified that fulfills all properties expected of a protein connecting actin filaments to cytoplasmic tails of adhesins. GAC, or glideosome-associated connector, binds and stabilizes F-actin at one end and at the other has a pleckstrin homology domain binding phosphatidic acid (PA) on the inner leaflet of the plasma membrane (Figure 3B). GAC can directly bind to cytoplasmic tails of adhesins. This has been shown for the adhesin MIC2 but has yet to be shown for any Plasmodium cytoplasmic tail such as those within EBA, PfRh, or AMA1. Interestingly, recent work in T. gondii has suggested that PA production is required for microneme secretion, suggesting that this phospholipid plays an important role in invasion (Bullen et al., 2016) (Figure 3B). This highlighted that the apical tip of zoites is a site of glideosome regulation. Of particular interest is that formin1 only localizes here, suggesting that filamentous actin is formed at the very tip. Furthermore, apical lysine methyl transferase is required for GAC to localize to the apical tip (Figure 3B). Therefore, it is likely that a signaling event received at the apical tip activates formin1 to produce F-actin while also releasing GAC to then associate with adhesins, subsequently driving motility (Figure 3B). Clearly there is still much to learn about this process.

Erythrocyte Heterogeneity, Parasite Invasion Ligand Polymorphism, and Plasmodium spp. Invasion

While much has been elucidated with regard to the core machinery driving invasion, the ligand-receptor interactions between the parasite and the host erythrocyte demonstrate considerable molecular heterogeneity and evolution. Polymorphism in parasite ligands allows the parasite to colonize diverse ecological niches as well as to evade the immune system. Conversely, host erythrocyte variation can serve to restrict the efficiency of Plasmodium invasion, a factor in limiting infections and pathogenesis.

Plasmodium spp. and Tropism for Their Vertebrate Hosts

A striking feature of Plasmodium spp. parasites is their vast distribution throughout vertebrates. The malaria parasites that infect primates appear to be quite limited in the number of species they can infect. For instance, the six malaria parasites infecting humans are rarely found in and poorly infect other Old World primates. Of great public health importance is the potential for cross-species transmission by Plasmodium spp. P. knowlesi is a zoonotic infection in Southeast Asia, transmitted to humans from macaques (Singh et al., 2004). Strikingly, it has recently been found that P. knowlesi populations in macaques fall into at least two isolated groups in pig-tailed and long-tailed macaques that can infect humans (Assefa et al., 2015). It is not clear whether mixing in humans has occurred, which could cause the formation of novel combinations and forms of invasion ligands with altered tropism, resulting in new patterns of virulence and transmission. P. falciparum has been occasionally found as an anthroponotic infection of bonobos (Krief et al., 2010; Liu et al., 2010). New World monkeys appear to be susceptible to P. falciparum and P. vivax, with the P. vivax-like parasite, P. simium, being found in natural infections of New World monkeys.

Recent studies shed considerable light on the origin of P. falciparum and P. vivax infections in human populations (Liu et al., 2010). Through the study of chimpanzee and gorilla feces, it has been shown these apes are commonly infected by Laveranian parasites closely related to P. falciparum. It appears P. falciparum has jumped to the human population from gorillas. Similarly, P. vivax-like parasites have been identified in chimpanzees and gorillas. Significant transmission appears to occur between these higher apes. However, all P. vivax infecting humans falls within the diversity of the chimp and gorilla P. vivax, suggesting that a rare event led to emergence of P. vivax in the human population. It is likely that invasion ligands have changed between the human P. falciparum and P. vivax and their ape counterparts to obtain optimal growth.

Most studies have focused on the various members of the RBL and EBL families as mediators of host cell tropism. The PfRh5 protein has recently become implicated in this host switch (Larremore et al., 2015), with the major change between Praefalciparum, the gorilla parasite which gave rise to P. falciparum, undergoing a remarkable horizontal transfer of a short region on chromosome 4, which encodes at least two important invasion genes, PfRh5 and PfRh4, from other Laveranian parasites, presumably resulting in the new erythrocyte specificity. PfRh5 was previously implicated in determining host specificity for human versus Aotus erythrocytes (Hayton et al., 2008). In addition, it was found that PfRh5 from P. falciparum bound chimp basigin more weakly than human and gorilla basigin (Wanaguru et al., 2013), mapping this difference in tropism to specific amino acid residues in this protein. The importance of PfRh5 in host cell tropism reflects the mechanistic importance of efficiency at this step in the invasion process, perhaps defining the first point of irreversible commitment to invasion by the merozoite.

In addition, erythrocyte sialic acid, often the terminal sugar on surface carbohydrates, is a key binding determinant of many parasite ligands. In P. falciparum there is a striking division between sialic acid-binding ligands, such as EBA-175, EBA-140, EBA-181, and PfRh1, and sialic acid-independent ligands such as PfRh2b, PfRh4, AMA-1, and MSP-1. It is possible that this dichotomy in invasion pathways evolved in response to the evolution in sialic acid binding. The enzyme CMAH catalyzes conversion of Neu5Ac to Neu5Gc, and is strikingly lost in humans and New World monkeys. Using transgenic methods, it was demonstrated that the presence of Neu5AC on human erythrocytes restricts the invasion of P. knowlesi (Dankwa et al., 2016). This suggests the Neu5AC/Neu5GC dichotomy in sialic acid has been a major restriction factor in the evolution of Plasmodium spp., as it has been for numerous other pathogens (Varki and Gagneux, 2009).

Blood Groups and Restrictions

Many receptors for parasite entry belong to human blood groups, and these proteins demonstrate considerable structural and expression-level polymorphisms. This includes the glycophorins, where significant polymorphisms and rearrangements are common in GPA, GPB, and GPC. Indeed, the GPA/GPB locus is significantly associated with severe malaria, suggesting invasion as a key determinant in pathogenesis (Malaria Genomic Epidemiology Network et al., 2015). CR1, which is bound by PfRh4 (Tham et al., 2010), harbors polymorphisms that vary its expression level with consequences in terms of P. falciparum invasion efficiency. Recently, basigin and DAF/CD55 have been identified as essential entry receptors for P. falciparum invasion. Sequence polymorphism identified in these receptors may be related to invasion efficiency. The basigin polymorphism that defines the Ok blood group has been associated with the efficiency of binding of the PfRh5 ligand (Crosnier et al., 2011). However, the major polymorphism of CD55 found in African populations is not associated with disease severity (Schuldt et al., 2017).

P. vivax is increasingly recognized as a major impediment to malaria elimination due to its ability to form hypnozoites in the liver. Innate resistance to P. vivax in humans of African descent was associated to the absence of the Duffy receptor DARC (Miller et al., 1976), largely absent from sub-Saharan African populations. However, reports of P. vivax in Duffy-negative individuals throughout sub-Saharan Africa are now common (Culleton et al., 2008). A priority is the identification of alternative essential ligand-receptor interactions in P. vivax. The genome sequence and transcriptomics reveals that several members of the RBL gene family are possible candidates (Figure 3A). Interestingly, a second member of the DBP family, known as PvEBP, has been identified (Hester et al., 2013), which binds to Duffy-positive reticulocytes in a Duffy-independent manner (Ntumngia et al., 2013) (Figure 3A).

Erythrocyte Age and Pathogenesis

Erythrocyte heterogeneity occurs in the normal course of cellular aging within ~120 days in circulation. Many Plasmodium spp. demonstrate a preference for the invasion of the youngest of erythrocytes (first 1–2 days), known as reticulocytes, which usually make up less than 1% of all erythrocytes. The strictest reticulocyte preference exists for P. vivax, resulting in a much lower parasite density in human infections compared with the lethal P. falciparum, which is much less restricted and can achieve high parasitemias. The P. vivax parasite ligand PvRBP2 was identified as preferentially binding to reticulocytes, potentially involved in the restriction of P. vivax (Galinski et al., 1992). P. falciparum parasite isolates also express a preference for reticulocytes, although the physiological relevance of this is unclear. An association between a loss in erythrocyte selectivity and disease severity has been observed (Chotivanich et al., 2000), which might be associated with a reduced reticulocyte preference, while to date no specific invasion pathway defining this preference has been identified. Surprisingly, such an association did not hold up in Africa, suggesting region-specific differences in invasion pathway usage (Deans et al., 2007).

A similar ability to adapt to colonize an expanded host niche in human erythrocytes has been demonstrated for the macaque parasite P. knowlesi (Lim et al., 2013). This is of growing significance, with increased P. knowlesi transmission to humans now occurring in Southeast Asia. In experiments selecting for increased P. knowlesi proliferation in human erythrocytes (Lim et al., 2013; Moon et al., 2013), amplifications and deletions in members of the DBL and RBL families has been observed (Dankwa et al., 2016; Moon et al., 2016).

Future Perspectives

Our understanding of the molecular interactions that occur during the invasion of erythrocytes by Plasmodium spp. has grown exponentially over the last 10 years. Key essential molecular interactions have now been defined, resulting in renewed impetus for the development of a blood-stage vaccine (Douglas et al., 2015). In addition, several essential signal transduction enzymes have been identified. The identification of small-molecule inhibitors could provide significant lead compounds for drug development targeting parasite invasion. High-throughput genetic screens, such as in P. berghei with PlasmoGem (http://plasmogem.sanger.ac.uk/) and in T. gondii, are beginning to reveal a plethora of essential targets. A community effort to produce a similar resource for P. falciparum and the other Plasmodium spp. would be a welcome development. Such mutant libraries could be powerfully combined with other perturbations, such as small-molecule inhibition, to specifically identify their targets.

What are some of the new frontiers for the next 10 years? Invasion has rarely been viewed in the context of in vivo infections. The development of intravital imaging for following Plasmodium spp. invasion in rodent malaria infections could reveal where, when, and how invasion in vivo may differ from in vitro observations. While genetic analyses have identified and validated several key molecules, little is known beyond their essentiality and phenotypic effects. It is critical to now explore how these molecules work together in protein complexes, and to determine the structural basis of their functions (Chen et al., 2014, 2017; Favuzza et al., 2017; Gruszczyk et al., 2016; Wright et al., 2014). Host erythrocyte genetics in vitro has also become permissible through the use of hematopoietic stem cells, which will reveal the host contribution to malaria infections. We have focused on invasion in this review; however, egress of Plasmodium spp. parasites from erythrocytes is also rich in signal transduction, and is directly upstream of invasion and intricately linked to pathways required for successful invasion. How egress and invasion are coupled remains a key question. Surprisingly, the parasite’s internal Ca²⁺ store required for invasion is still uncertain, and the molecular mechanism by which Ca²⁺ is released remains unknown, while no obvious IP₃ receptors/ryanodine receptors are encoded within the parasite genome.

What do we need in terms of experimental systems? Most research has been carried out in in vitro cultured P. falciparum. The macaque parasite P. knowlesi can now be cultured in vitro in human erythrocytes, making it an accessible system (Lim et al., 2013; Moon et al., 2013). A high priority is the establishment of in vitro P. vivax culture. P. vivax has so far been accessible through the analysis of patient isolates or parasites from non-human primate infections. Of great interest will be comparative studies looking at the conserved and divergent mechanisms that occur between Plasmodium spp. parasites at different stages, and also those that occur between Plasmodium spp. and related apicomplexan parasites such as Toxoplasma, Cryptosporidium, and Babesia parasites. The harnessing and development of experimental systems to interrogate invasion in multiple species will pave the way toward an understanding of conserved and species-specific features.

Ten years ago, erythrocyte invasion was thought to be a very difficult process to study. Host cell invasion by apicomplexan parasites is a unique active process on which model organisms can shed little light, and will continue to engage and challenge researchers well into the next decade.