Plasmodium vivax Reticulocyte Binding Proteins for invasion into reticulocytes

Abstract Plasmodium vivax is responsible for most of the malaria infections outside Africa and is currently the predominant malaria parasite in countries under elimination programs. P. vivax preferentially enters young red cells called reticulocytes. Advances in understanding the molecular and cellular mechanisms of entry are hampered by the inability to grow large numbers of P. vivax parasites in a long‐term in vitro culture. Recent progress in understanding the biology of the P. vivax Reticulocyte Binding Protein (PvRBPs) family of invasion ligands has led to the identification of a new invasion pathway into reticulocytes, an understanding of their structural architecture and PvRBPs as targets of the protective immune response to P. vivax infection. This review summarises current knowledge on the role of reticulocytes in P. vivax infection, the function of the PvRBP family of proteins in generating an immune response in human populations, and the characterization of anti‐PvRBP antibodies in blocking parasite invasion.

and will generally invade reticulocytes. Since the establishment of a continuous in vitro culture for P. falciparum in the late 1970s, the field of malaria parasite invasion has been dominated by studies of P. falciparum invading normocytes. Collectively, these studies have provided insights into the step-wise nature of parasite entry, have identified parasite and host factors involved in invasion, and led to the development of inhibitors and antibodies that can block parasite invasion and provide protection from clinical disease (Paul et al., 2015;Cowman et al., 2017;Draper et al., 2018). In contrast, P. vivax invasion into reticulocytes is poorly understood due to the lack of a long-term in vitro culture system for this parasite species (Kanjee, Rangel, Clark, & Duraisingh, 2018;Tham et al., 2017).
The preference of P. vivax for reticulocytes has implications in infection dynamics, parasite reservoirs, and potential parasite killing mechanisms. There are two distinct classes of reticulocytes that are present within the bone marrow compartment and in peripheral circulation (Griffiths et al., 2012). In the bone marrow compartment, R1 reticulocytes that have expelled the nucleus, but retain residual reticulum and are motile and multi-lobular. R2 reticulocytes are released from the bone marrow to the peripheral circulation and are non-motile and mechanically stable. As these reticulocytes mature in the bone marrow and in peripheral circulation, they remove all their organelles and lose 20% of their plasma membrane surface area (Moras, Lefevre, & Ostuni, 2017). Reticulocytes express several surface proteins that are lost as they mature into normocytes. In particular, CD71 (Transferrin Receptor 1, TfR1), CD49d, CD151, CD81, and CD82 are present only on young reticulocytes compared with mature red blood cells . Using short-term ex vivo cultures, P. vivax has been observed to have higher invasion rates into reticulocytes with high levels of TfR1 compared with reticulocytes with lower levels of TfR1 (Malleret et al., 2014). In the same study, P. vivax invasion into TfR1 high-reticulocytes caused a more rapid loss of TfR1 and expulsion of the residual reticulum compared with uninfected reticulocytes. However, a study using Indian P. vivax strains showed large differences in reticulocyte preferences . Although there was a low prevalence of circulating schizonts (the mature replicative form of the parasite), there was an association between increased reticulocyte preference and the number of schizonts, suggesting a potential link between invasion of younger reticulocytes and effective parasite development.
This study also showed the detection of early-stage P. vivax infection in reticulocytes with visible reticulum staining, suggesting that modifications to reticulocytes as observed ex vivo may not happen as rapidly in vivo . In a separate study, it was also shown that P. vivax had normal growth and development in TfR1-high reticulocytes in G6PD-Mahidol mutants suggesting an advantage to invasion of reticulocytes in these settings (Bancone et al., 2017).
Full-length genes encode large molecular weight proteins of over 280 kDa. Almost all PvRBPs have a signal peptide at the N-terminus and a putative transmembrane domain at the C-terminus. Transcriptome analyses show that several of the PvRBPs are expressed in field isolates (Bozdech et al., 2008).
The first functional study on PvRBPs showed that native PvRBP1a and PvRBP2c are expressed at the apical tip of P. vivax merozoites and form a high molecular weight complex that binds reticulocytes (Galinski, Medina, Ingravallo, & Barnwell, 1992). Recent studies using recombinant PvRBP proteins have further described their binding characteristics and defined regions of the proteins involved in binding red blood cells (summarised in Table 1). In particular, PvRBP1a has been particularly well characterized and is shown to bind preferentially to reticulocytes (Franca et al., 2016;Gupta et al., 2017;Han et al., 2016;Ntumngia et al., 2018). Although several studies have purified different PvRBP1a fragments, the collective results show that two of the best binding fragments range from residues 30-778 and 157-650 (Gupta et al., 2017;Ntumngia et al., 2018). In contrast, it has also been shown that PvRBP1a (residues 160-1170) binds normocytes (Franca et al., 2016). These conflicting results may indicate that in comparison with shorter recombinant fragments, PvRBP1a (residues 160-1170) may not be folded correctly and therefore not reflect its true binding properties. Recombinant PvRBP1b (residues 339-587), PvRBP2c (residues 464-876), and PvRBP2b (residues 168-1124) also showed preferential binding to reticulocytes (Franca et al., 2016;Gupta et al., 2017;Han et al., 2016), whereas recombinant PvRBP2a (residues 160-1135) has been shown to bind to both normocytes and reticulocytes (Franca et al., 2016).
The use of recombinant PvRBP proteins have facilitated the determination of their functional properties; however, it would be important to examine if the respective native parasite proteins also display similar binding characteristics. Gupta and colleagues have also determined that the native PvRBP1a and PvRBP2c demonstrate the same binding profiles as has been previously shown by Galinski et al. (1992, Gupta et al., 2017. Experiments with native parasite proteins are extremely challenging for P. vivax research due to the absence of a long-term in vitro culture system and limited access to clinical P. vivax samples. Nevertheless, it will be crucial to validate the binding profiles obtained with recombinant proteins with what is observed by using parasite lysates in which the PvRBP proteins are folded and functional for invasion.

| NEW INVASION PATHWAY INTO RETICULOCYTES
Since the 1970s, it has been known that P. vivax invasion into red blood cells is dependent on Duffy Antigen Receptor for Chemokine (DARC;Miller, Mason, Clyde, & McGinniss, 1976). DARC is recognised by P. vivax Duffy Binding Protein (PvDBP) to mediate an essential step in P. vivax invasion, and many studies have contributed to the development of PvDBP as the lead P. vivax vaccine candidate, including the characterisation of anti-PvDBP inhibitory antibodies that block invasion. However, DARC is present on both normocytes and reticulocytes and therefore cannot govern specific recognition of reticulocytes by P. vivax. Although it has been proposed that there is exposure of PvDBP binding site on DARC in young reticulocytes that allows it to selectively bind reticulocytes, this mechanism remains to be tested further by other laboratories (Ovchynnikova et al., 2017).
Recent studies show that Transferrin Receptor 1 (TfR1, CD71) binds to PvRBP2b to mediate a critical pathway into reticulocytes Gruszczyk, Kanjee, et al., 2018). TfR1 binds to its human ligand, iron-loaded transferrin (Tf), and this complex regulates one of the main mechanisms for transporting iron into cells (Kawabata, 2019). TfR1 is highly expressed on reticulocytes and is selectively lost as they mature into normocytes. TfR1 mutant cells that are deficient in TfR1 expression have a strong defect in P. vivax invasion, showing that TfR1 mediates a critical invasion pathway into reticulocytes (Gruszczyk, Kanjee, et al., 2018). Mouse monoclonal antibodies against PvRBP2b also inhibit PvRBP2b binding to reticulocytes and block complex formation with TfR1. These anti-PvRBP2b monoclonal antibodies also resulted in a reduction in parasite invasion using Thai and Brazilian clinical isolates (Gruszczyk, Kanjee, et al., 2018).
A 3.7 Å cryo-electron microscopy (cryo-EM) structure of the ternary complex of PvRBP2b, TfR1, and Tf has shed light on the critical residues involved in mediating complex formation (Gruszczyk, Huang, et al., 2018). The ternary complex is composed of homodimeric TfR1 (residues 120-760) bound to two molecules of iron-loaded Tf (residues 1-679), with two molecules of PvRBP2b (residues 168-633) bound on either side. The most extensive interaction site is between PvRBP2b and TfR1, which has a surface buried area of ~1271 Å 2 , whereas the interaction site with Tf has a surface buried area of~386 Å 2 . The region of PvRBP2b from 169 to 470 is highly polymorphic and under balancing selection (Gruszczyk, Huang, et al., 2018). Although this region binds to TfR1 and Tf, none of the prevalent field polymorphisms map to the amino acid residues on PvRBP2b that form critical contacts with either TfR1 or Tf as observed by the cryo-EM structure ( Figure 1). Extensive mutagenesis experiments identified three critical pairs of interacting residues for PvRBP2b-TfR1 complex formation: those formed by PvRBP2b (Y542) and TfR1(Y211), a salt bridge formed between PvRBP2b(K600) and TfR1(E294), and a second salt bridge between TfR1(E149) and PvRBP2b(R359) (Figure 1).
At present, apart from TfR1, there has been no additional reticulocyte receptors that have been implicated as cognate receptors for PvRBPs. PvRBP1a binding is sialic acid independent but trypsin and chymotrypsin sensitive , whereas PvRBP2c binding is insensitive to all three enzyme treatments (Gupta et al., 2017). These results strongly suggest that other reticulocyte receptors will be important for P. vivax invasion

| TfR1 AS A CELLULAR RECEPTOR FOR NEW WORLD HAEMORRHAGIC FEVER ARENAVIRUSES
TfR1 is also a cellular receptor for human New World haemorrhagic fever arenaviruses, including Machupo (MACV), Junin, Guanarito, and Sabiá viruses (Abraham, Corbett, Farzan, Choe, & Harrison, 2010;Radoshitzky et al., 2007). Residues 208 to 212 of the TfR1 apical domain provide a critical recognition site for these viruses (Abraham et al., 2010). Of particular interest, Y211 that is localized in the apical domain of TfR1 is a critical residue for entry of these New World haemorrhagic arenaviruses and P. vivax (Abraham et al., 2010;. Furthermore, the binding of recombinant PvRBP2b can compete with the binding of recombinant MACV Glycoprotein 1 to TfR1, showing that these two pathogens have co-opted a similar site on the apical domain of TfR1 for entry into their target cells (Gruszczyk, Kanjee, et al., 2018).

| STRUCTURAL SCAFFOLDS OF PvRBPs
The first description of a family of Reticulocyte or Normocyte binding proteins (RBP or NBP) was described in P. yoelli as the Py235 proteins (reviewed in Gruner et al., 2004). The description of the Py235 and the PvRBP family led to the identification of the homologous P.
PfRh5 cannot be deleted in any strain (Baum et al., 2009;Hayton et al., 2008), suggesting that it would be an ideal component targeting the blood stages in an effective P. falciparum vaccine. PfRh5 is smaller than the other PfRh members and does not contain a transmembrane region (Baum et al., 2009;Hayton et al., 2008).  PfRipr/CyRPA/PfRh5-Basigin complex is required for triggering the release of Ca 2+ and establishing the tight junction (Volz et al., 2016).
These observations show that the PfRh5/PfRipr/CyRPA complex is essential in the sequential molecular events leading to P. falciparum invasion of human red blood cells.
In P. vivax and P. knowlesi, the exact homologue to PfRh5 remains elusive, but the homologs for PvRipr and PvCyRPA have been identified (Hoo et al., 2016). In a recent study, P. knowlesi PkRipr and PkCyRPA were shown to be essential for parasite viability (Knuepfer et al., 2019). Furthermore, PkRipr did not form a complex with PkCyRPA, but instead formed a trimeric complex with thrombospondin-related anonymous protein (PkPTRAMP) and cysteine-rich small secreted protein ( TfR1 and PfRh5 to Basigin (Crosnier et al., 2011;Gruszczyk, Kanjee, et al., 2018;Wright, Hjerrild, Bartlett, Douglas, et al., 2014a). In contrast to PfRh5 that receptor recognition site is located on the tip of the domain, binding of PvRBP2b to TfR1-Tf involves residues located at the side of the domain and residues of the adjacent α-helical domain of this protein. The receptor for PvRBP2a is currently unknown.
The surface charge distribution of PvRBP2a, PvRBP2b, and PfRh5 is different (Figure 2). In particular, PvRBP2a has a distinctive negative patch on the apex of the molecule that corresponds to the Basiginbinding area of PfRh5, whereas PfRh5 as well as the N-terminal domain of PvRBP2b mostly displays a positive charged surface (Gruszczyk, Kanjee, et al., 2018;Wright, Hjerrild, Bartlett, Douglas, et al., 2014a). The differences in surface charge may be important in mediating the specific interactions between these parasite adhesins and their respective red blood cell receptors.

| ANTIBODIES TO PvRBP2b AND PfRh5
Antibodies Similar to anti-PfRh5 antibody QA1, antibody 10B12 engages its antigen at the tip of the N terminal domain, probably also facing the loop which is stabilised by a disulfide bond .
The presence of the corresponding disulfide bridge and loop of PvRBP2a is also required for the protein function as mutation of C299 and C303 to serine residues results in a loss of the capability of PvRBP2a to bind erythrocytes while not altering the overall structural conformation of the protein .
Collectively, these results highlight the importance of the disulfide

| NATURALLY ACQUIRED ANTIBODY RESPONSE TO PvRBPs
The study of the human immune response to malaria infection is central to the discovery of new vaccine candidates and biological markers of disease. The availability of the P. vivax parasite genome since 2008 has enabled researchers to examine a much larger panel of proteins (Baum et al., 2016;Finney et al., 2014;Franca et al., 2017;Hostetler et al., 2015;, not least the entire family of PvRBPs. Although these studies show that many other parasite antigens that are associated with protection, we will only focus on the results for PvRBPs for this review. To date, all PvRBPs have been subjected to serological examinations and are all shown to be targets of natural antibody responses (Franca et al., 2016;. Plasma samples from different endemic regions have been used to study natural acquisition of antibodies to PvRBPs, from the relatively high transmission areas in India (Gupta et al., 2017) and Papua New Guinea (PNG; Franca et al., 2016;Ntumngia et al., 2018) to the lower transmission sites in Brazil (Ferreira et al., 2014;Longley et al., 2018;Ntumngia et al., 2018;Tran et al., 2005), Cambodia (Hostetler et al., 2015), Colombia (Rojas Caraballo, Delgado, Rodriguez, & Patarroyo, 2007), Korea (Han et al., 2015;Han et al., 2016), Solomon Islands (Longley et al., 2018), and Thailand (Hietanen et al., 2015;Longley et al., 2018). These samples encompassed specimens from uninfected endemic residents, infected asymptomatic P. vivax carriers, and acute P. vivax patients. The diversity of these samples provides rich data for identifying new vaccine candidates as well as serological markers of P. vivax malaria.
PvRBP1a was the first protein in the PvRBP family whose antibody response was systematically examined . In this initial study published in 2005, plasma specimens from the Brazilian Amazon state of Rondonia were tested against five recombinant PvRBP1a fragments, which together span nearly the entire extracellular domain (>2,600 amino acids). The study revealed key features of the natural antibody response that appear to be common among PvRBPs, including (a) positive correlation between the antibody level and parasite exposure, (b) IgG subtypes being biased towards cytophilic IgG1 and IgG3,and (c) association between the antibody level and clinical protection. To date, PvRBP1a is the only full-length PvRBP whose entire extracellular sequence has been mapped serologically.
Comparison of the IgG responses across different subdomains of PvRBP1a revealed stronger reactivity towards the N-terminal half, a finding which was later confirmed in a separate study . This N-terminal bias coincides with a higher degree of genetic diversity of the N-terminal domain (Rayner et al., 2005), suggesting a stronger immunogenic selection pressure on this region of the protein.
In the endemic populations, the magnitudes of IgG responses to different PvRBPs are generally correlated (Gupta et al., 2017;Hietanen et al., 2015) and tend to increase with age (Hietanen et al., 2015;. This co-acquisition of antibodies to different PvRBPs is unlikely due to cross-reactivity given the limited sequence similarity between these proteins. Not surprisingly, human IgG obtained by affinity purification with PvRBP1a showed little cross-reactivity against PvRBP2c, and vice-versa (Gupta et al., 2017).
Rabbit immunizations with recombinant PvRBPs also generated antibodies specific to the respective antigen (Franca et al., 2016).
Therefore, the co-development of IgG responses to different PvRBPs is likely due to repeated P. vivax infections that expose the human body to all proteins simultaneously. In addition to age, other measures of parasite exposure such as the number of previous malaria episodes , years of residence in endemic areas , and the estimated number of new infections acquired in the lifetime (Franca et al., 2016) have been shown to positively associate with anti-PvRBP reactivities.

| SEROLOGICAL MARKERS OF EXPOSURE
People with concurrent P. vivax infection, both symptomatic and asymptomatic, tend to have higher levels and breadth of IgG responses to PvRBPs than do uninfected individuals (Franca et al., 2017;Longley, França, et al., 2017). The heightened sero-reactivities in infected individuals suggest that PvRBPs may be viable markers for comparing malaria burdens in endemic populations. Indeed, the IgG responses to PvRBP1a and PvRBP2c fragments have been shown to map spatial heterogeneity in western Thailand  and be able to differentiate different endemic communities in Brazilian Amazon . The magnitudes of IgG responses to several PvRBPs were also found to decline after clearance of infection, with estimated half-lives of 2-18 months .
Therefore, antibodies to some of these proteins, especially the shortlived ones, may be good surrogates of recent exposure. In a recent attempt to develop a panel of serological markers of recent infections (Longley et al., 2018), a recombinant fragment of PvRBP2b was the best performing antigen among the >300 P. vivax proteins tested.
The IgG response to this protein has the sensitivity and selectivity of 74% in detecting P. vivax infection within the previous 9 months.
Combining PvRBP2b with other proteins further improved the diagnostic performance, making antibody-based surveillance an attractive approach for malaria control and elimination programs (Greenhouse et al., 2019).

| SEROLOGICAL MARKERS OF PROTECTION
Identification of parasite proteins associated with clinical protection is an approach often taken in malaria vaccine discovery (Cutts et al., 2014;Franca et al., 2017). Because naturally acquired immunity against malaria is gained through repeated infections, an association between the antibody level and protection is often confounded by exposure, a factor that has to be accounted for in sero-epidemiological analysis. The most comprehensive study to examine association between the IgG reactivities to PvRBPs and clinical protection (Franca et al., 2016) came from a cohort of young children in East Sepik, PNG, whose exposure to P. vivax was defined by molecular force of blood stage infection ( mol FOB, the incidence of new blood infection). In this study, children 1-3 years old were followed for up to 16 months, with blood sampling every 8 weeks and at clinical malaria episodes. P. vivax infections were detected by PCR, and the parasites were genotyped for determination of mol FOB. After adjusting for exposure, IgG for each of the six PvRBPs tested (PvRBP1a, PvRBP1b, PvRBP2a, PvRBP2b, PvRBP2c, and PvRBP2p2) was associated with protection, with a reduction of 30-50% in risk of P. vivax malaria observed in children who had medium to high antibody levels. However, in multivariate analysis to account for co-acquisition of antibodies to different PvRBPs, only IgG to PvRBP1a and PvRBP2b remained, with IgG1 being the underlying antibody subclass associated with protection. Notably, these two proteins were amongst the strongest protective antigens in the panel of 38 P. vivax antigens tested with the same methodology (Franca et al., 2017). Nevertheless, it is important to note that these studies were only performed with recombinant proteins, which may have inherent caveats such as the use of a single variant and that the protein folding may be different from native protein. In particular, for proteins that are highly polymorphic, immune responses to a different variant may be very different. Consistently, the magnitude of IgG to PvRBP1a fragments had previously been associated with duration since the last known malaria episode, a proxy of protection, in a Brazilian cohort . The magnitude of IgG response to PvRBP2b had also been associated with reduced parasitaemia in Thai adult patients independently of age (Hietanen et al., 2015). Thus, among the PvRBPs, PvRBP1a and PvRBP2b appear to be the strongest vaccine candidates.

| CONCLUSIONS
The PvRBP family of P. vivax parasite adhesins is clearly emerging to be important for reticulocyte invasion and is the target of naturally acquired immunity in many different transmission settings. Future studies on the identification of other reticulocyte receptors for PvRBP and the characterization of inhibitory antibodies will provide insightful information on the multiple invasion pathways of P. vivax and how to target them to block entry into reticulocytes.