Vγ9vδ2 T Cells Kill Intraerythrocytic P. Falciparum in a Granzyme B Dependent Manner

Vγ9Vδ2 T Cells Kill Intraerythrocytic P. Falciparum in a Granzyme B Dependent Manner

The Harvard community has made this article openly available. Please share how this access benefits you. Your story matters

Citation LIANG, ZHITAO. 2019. Vγ9Vδ2 T Cells Kill Intraerythrocytic P. Falciparum in a Granzyme B Dependent Manner. Master's thesis, Harvard Medical School.

Citable link http://nrs.harvard.edu/urn-3:HUL.InstRepos:42057400

Terms of Use This article was downloaded from Harvard University’s DASH repository, and is made available under the terms and conditions applicable to Other Posted Material, as set forth at http:// nrs.harvard.edu/urn-3:HUL.InstRepos:dash.current.terms-of- use#LAA

Vγ9Vδ2 T Cells Kill Intraerythrocytic P. falciparum in a Granzyme B Dependent Manner

Zhitao Liang

A Thesis Submitted to the Faculty of

The Harvard Medical School

in Partial Fulfillment of the Requirements

for the Degree of Master of Medical Sciences in Immunology

Harvard University

Boston, Massachusetts.

May, 2019

Thesis Advisor: Dr. Judy Lieberman & Dr. Caroline Junqueira Zhitao Liang

Vγ9Vδ2 T Cells Kill Intraerythrocytic P. falciparum In a Granzyme B Dependent Manner

Abstract

Malaria, caused by Plasmodium falciparum, is one of the most serious infectious diseases and a leading cause of morbidity in the developing world. Humoral immunity against blood stage malaria has been studied for decades. Now emerging evidence suggests that cytotoxic cells also contribute to the defense against intraerythrocytic parasites. P. falciparum infects mature red blood cells (RBCs) which lack classical HLA expression, thereby escaping from the surveillance of conventional T cells. P. falciparum infected patients show an expansion of non-MHC restricted innate-like Vγ9Vδ2 T cells and elevation of the antimicrobial protein granulysin in their blood, suggesting this subset of killer cells might respond to malaria infection. However, whether Vγ9Vδ2

T cells kill intraerythrocytic P. falciparum remains unknown. We found that Vγ9Vδ2 T cells release their cytotoxic granules when exposed to P. falciparum-infected RBCs. Both Vγ9Vδ2 T cells and purified granulysin and granzyme B lyse infected RBCs, suggesting that Vγ9Vδ2 T cells kill the intraerythrocytic parasites in a granulysin and granzyme B dependent manner. Proteomics analysis to identify GzmB substrates shows that GzmB disrupts biosynthetic and metabolic pathways that are essential for the survival of P. falciparum.

Table of Contents

1. Chapter 1: Background & Introduction

1.1 Overview

1.2 Malaria Parasite Life Cycle

1.3 Human Immune Response to Erythrocytic Stage P. falciparum

1.3.1 Humoral Immunity Restricts the Spread of Parasites

1.3.2 Cellular Immunity and Evasion Mechanism of P falciparum

1.4 Role of Cytotoxic Cells in the Immune Response Against Blood Stage P. falciparum

1.4.1 Cytotoxic Lymphocytes and Killing Mechanisms

1.4.2 NK cells

1.4.3 Vγ9Vδ2 T cells

1.5 Research Goals

2. Chapter 2: Results and Methods

2.1 Results

2.2 Materials and Methods

3. Chapter 3: Discussion and Perspectives

4. Bibliography

iii

List of Figures

Figure 1: life cycle of P. falciparum

Figure 2: granzymes and their targets

Figure 3: phosphoantigen biosynthesis and γδ TCR ligands

Figure 4: Vγ9Vδ2 T cells target infected erythrocytes in a contact dependent manner

Figure 5: purified GNLY & GzmB kill intraerythrocytic P. falciparum

Figure 6: candidate GzmB substrates in P. falciparum lysate

Figure 7: 2D gel electrophoresis identification of substrates in P. falciparum lysate

Figure 8: GzmB cleaves P. falciparum enolase and disrupts glycolysis

Figure 9: GzmB disrupts the purine salvage pathway

Figure 10: putative GzmB substrates in crude P. falciparum mitochondrial fractions

Figure 11: GNLY and GzmB disrupts P. falciparum mitochondrial function

Figure 12: putative GzmB substrates and targeted pathways

Acknowledgements

I am tremendously thankful to Prof. Judy Lieberman for her support and mentorship. It is my pleasure and privilege to work under her supervision for the past one year. I admire her talents and insights into cutting-edge science, and I have learnt how scientific work should be performed.

I am incredibly grateful to Dr. Caroline Junqueira for her close supervision and training throughout my time in the lab. I have learnt a lot regarding experimental design, data analysis and paper writing by working together with Dr. Caroline Junqueira. I appreciate her kindness for supporting me on every experiment and openness to my questions.

I appreciate other members in the lab, especially Dr. Sumit Sen Santara, who offered training and help for biochemistry aspects of the thesis. I also want to acknowledge the faculty and staff of the

Master Program in Immunology.

This work was conducted with support from Students in the Master of Medical Sciences in

Immunology program of Harvard Medical School. The content is solely the responsibility of the author and does not necessarily represent the official views of Harvard University and its affiliated academic health care centers.

Statement of Research

The work in this thesis was performed under the direction of Prof. Judy Lieberman and Prof.

Caroline Junqueira in the Program in Cellular and Molecular Medicine, Boston Children’s Hospital.

All data provided in this thesis were generated by the author, except where noted. Figure 4 & 5 were produced by the author and Dr. Junqueira. All experiments presented here were performed between

Feb, 2018 and May, 2019.

Chapter 1: Background & Introduction

1.1 Overview

Malaria, a mosquito-borne parasitic disease caused by Plasmodium spp, is one the most serious infectious diseases worldwide. The transmission of malaria is usually regional, and it is typically distributed in tropical and subtropical countries. Despite the application of vector control, administration of anti-malarial drugs, and other protective measures against the mosquito vector, such as insecticide and mosquito nets, malaria still infects about 219 million people worldwide and claims the lives of more than 435,000 people each year according to the WHO World Malaria

Report of 2018 [1]. Around 80% of global malaria deaths were reported from 17 countries in Africa and India. People with malaria usually experience flu-like symptoms, including fever, chills, headache, nausea, vomiting, muscle pain and fatigue. Symptoms caused by hemolysis, including anemia and jaundice, are observed around 10 to 15 days after infection. If untreated, severe clinical manifestations of malaria, such as kidney failure, seizures, coma, and even death occur. Children under 5 years old are the most vulnerable group and account for more than 60% of malaria-caused deaths worldwide. Treatment with anti-malarial drugs, such as quinolines, naphthoquinones, antifolates, and artemisinin(ART)/artemisinin-based combination therapy (ACTs) have reduced case incidence and global burden of malaria since 2000 [1].

Of more than 120 Plasmodium species that infect vertebrates, only six strains highly infect humans. P. knowlesi, P. malariae and P. ovale have low incidence and are sensitive to chloroquine and ACTs [1]. P. vivax, the predominant parasite in America, has shown increased resistance to chloroquine [2]. P. falciparum, the most prevalent malaria parasite in Africa, is the most drug- resistant Plasmodium species that infects humans and is responsible for the majority of malaria- related deaths. Resistance of P. falciparum to ART was first documented in 2008 and has been

reported throughout most of the Greater Mekong countries [3]. Resistances of parasites to ART derivatives and of the vector to insecticides are significant challenges to malaria control. Resistance to ART is potentially due to an altered capacity of the parasites to deal with cellular stress by modulating the unfolded protein response and the ubiquitin/proteasome system [4]. Mutations in the

Kelch-like protein K13, which leads to a reduction of phosphatidylinositol-3-kinase polyubiquitination, may also contribute to drug resistance of P. falciparum [5]. However, the exact mechanism of resistance remains largely unclear.

Despite extensive pre-clinical testing in animal models and more than 200 trials in humans with purified parasite immunogens, most vaccines have limited protective effects and have failed to show long-lasting benefits at a population level. The fusion of the S antigen of the hepatitis B virus with RTS, S, a chimeric molecule produced by the sporozoite surface circumsporozoite protein, is the only effective candidate when delivered with strong adjuvants. Although it has advanced to

Phase III clinical trials, its efficacy is still less than 50% [6]. Only vaccination with attenuated parasites blocks transmission [7]. The greatest obstacle to current development of effective vaccines against malaria is our incomplete understanding of the interaction between the immune system and malaria parasites. Thus, it is important to study in more detail the immune response against malaria infection. This thesis will focus on studying a novel immune response against P. falciparum.

1.2 Plasmodium falciparum Life Cycle

The malaria parasite life cycle can be divided into two stages: (1) the asexual stage with continuous cyclic proliferation, and (2) the sexual stage in which sporozoites are produced. When an infected female Anopheles mosquito bites an individual, ~100 sporozoites are transmitted from the vector to the skin of the host. It takes ~2 h for sporozoites to leave the skin and travel through the bloodstream to the liver [8], where they invade host hepatocytes using a cholesterol uptake

2 pathway [9], form a parasitophorous vacuole to avoid lysosomal degradation, and start the asexual replication cycle. After maturation to late schizonts, merozoites are released from infected hepatocytes into merosomes and enter the blood. Merozoites attach to erythrocytes with merozoite surface protein (MSP), align their apical ends with erythrocyte membranes using merozoite apical membrane antigen (AMA), form tight junctions via interactions between erythrocyte-binding ligands (EBLs) and reticulocyte binding-like protein homologs (RBLs), invade RBCs and initiate blood-stage asexual replication [10]. The asexual cycle of P. falciparum takes around 48 h, starting from post-invasion ring stage to mature trophozoite stage, and then to schizonts (Figure 1). Clinical symptoms coincide with the rupture of RBCs when the released mature merozoites invade new

RBCs [11].

A subpopulation of parasites leaves the bloodstream and initiates sexual development to produce female and male gametocytes. Bone marrow is the primary location of gametocyte maturation, but some immature gametocytes are also observed in the spleen. Gametocytes re-enter the peripheral circulation and are then ingested by a mosquito. Gametocytes undergo gametogenesis in the midgut of the mosquito and become microgametes (male gametes) and macrogametes

(female gametes) and undergo fertilization to form zygotes. Around 5 to 10% of zygotes differentiate into ookinetes, of which a small portion develops into oocysts. Ookinetes invade the midgut epithelium via lectin binding, and oocysts produce sporozoites that later migrate to the salivary glands of the mosquito, where they are transmitted to a new host and re-start the cycle.

This thesis will focus on understanding the host immune response to erythrocytic stage P. falciparum.

Figure 1: life cycle of Plasmodium falciparum [12] Malaria parasites are transmitted as sporozoites when an infected female Anopheles mosquito takes a bite of the human host. Sporozoites in the blood are collected in the liver, where they invade hepatocytes and start asexual replication. Liver stage replication takes around seven days and ~40,000 merozoites leave hepatocytes and enter the peripheral blood. The released merozoites invade circulating RBCs and begin asexual replication. It takes ~48 hours for the parasites to replicate and mature from the ring stage to the trophozoite stage and to the schizont stage. At late schizont stage, RBCs rupture and release 8 to 32 daughter merozoites into the circulation to initiate another asexual replication cycle. Besides the asexual cycle, a small subset of parasites leaves the blood, enters the extravascular space of the bone marrow, undergoes sexual maturation through I-V stages and produce sexual gametocytes. The resulting gametocytes re-enter the peripheral circulation and are ingested by a mosquito. Gametocytes of opposite sex rapidly undergo gametogenesis in the midgut of the mosquito and become microgametes (male gametes) and macrogametes (female gametes). Zygotes produced by the fertilization of macrogametes develop into invasive ookinetes, which then penetrate the mosquito gut wall, form oocysts and undergo asexual replication to produce sporozoites. Sporozoites migrate to the salivary glands after the rupture of the oocyst and wait for a new round of transmission from mosquito to human host.

1.3 Human Immune Response to Erythrocytic Stage P. falciparum

1.3.1 Humoral Immunity Restricts the Spread of Parasites

The immune response against blood-stage malaria infection depends highly on opsonizing antibodies specific to merozoite surface antigens that provide protection against erythrocytic stage parasite infection by blocking either the invasion or emergence of merozoites, and facilitating the clearance of infected red blood cells by Antibody-Dependent Cellular Cytotoxicity (ADCC)/

Phagocytosis (ADCP) and Inhibition (ADCI) [13]. Besides opsonizing antibodies, neutralizing antibodies are produced against parasite toxins, such as malaria pigment and by-products of heme degradation during parasite metabolism [14]. P. falciparum-infected RBCs escape from splenic clearance by adhering to various small and medium-sized blood vessels via erythrocyte membrane protein 1 (PfEMP1). Sequestration of infected RBCs leads causes blood vessel injury and microvascular obstruction [15], which is responsible for some of the severe symptoms in patients.

Antibodies against such cytoadhering receptors promote the phagocytosis and destruction of infected RBCs [16]. The malaria antibody response is short lived, and declines within 3 to 9 months post infection [17].

1.3.2 Cellular Immunity and Evasion of P. falciparum

CD4 T cells are involved in humoral immunity as cytokines released by CD4 T cells influence the isotype of antibodies that are produced. The virulence of P. falciparum is highly associated with evasion of host immune responses: parasites utilize antigenic variation, polymorphism and sequestration to escape from antibody recognition. In chronic infection, intraerythrocytic parasites dampen the immune response by promoting the growth of FOXP3-CD45RO+CD4+ helper T cells, which produce IL10 [18] and induce the B and T cell exhaustion [19]. IFN-γ secreting CD8 T cells might prevent the development of chronic infection [20]. Monocytes and neutrophils phagocytose

5 infected RBCs mainly via the interaction between CD36 and PfEMP-1 [21] and secrete cytokines that stimulate adaptive immunity. Macrophages kill gametocytes by nitric oxide production.

However, both phagocytosis and radical oxygen species generation are impaired when phagocytes ingest malaria hemozoin [22].

1.4 Role of Cytotoxic Cells in Immune Response Against Blood-stage P. falciparum

Circulating CD8 T cells from infected patients recognize and kill P. vivax infected reticulocytes in an HLA-dependent manner [23]. Recognition and killing depends on the fact that reticulocytes retain protein translation machinery and express HLA class I. However, CD8 T cells cannot kill blood-stage P. falciparum because RBCs do not express MHC. This led us to ask whether MHC- independent cytotoxic cells contribute to defense against erythrocytic P. falciparum.

1.4.1 Cytotoxic Lymphocytes and Killing Mechanisms

Killer lymphocytes trigger programmed cell death in target cells by ligating and activating death receptors (e.g. FAS, TNFR1, DR4) or releasing their cytotoxic mediators at the immunological synapse. Cytotoxic granules contain granzymes (Gzms) and two pore-forming proteins: granulysin

(GNLY), which preferentially forms pores in cholesterol-poor microbial membranes, and perforin

(Pfn), which selectively penetrates cholesterol-containing membranes (Thiery & Lieberman, 2014).

Pfn disrupts target cell membrane integrity, initiates a membrane repair response, and delivers the

Gzms into the cell. Gzms cleave proteins involved in pathways that are essential for survival and activate multifocal programs of cell death that morphologically resemble apoptosis. Among the 5

Gzms (GzmA, B, H, K and M) in humans (Figure 2), GzmA and GzmB are the most abundant enzymes in killer lymphocytes. GzmA cleaves after Arg or Lys and GzmB cleaves after Asp residues. GzmA-mediated caspase-independent cell death is caused by single-stranded DNA nicks, loss of mitochondrial inner membrane potential and the release of reactive oxygen species (ROS),

6 which causes the translocation of the ER-associated SET complex to the nucleus [24]. GzmB induces caspase-dependent and independent apoptosis [25]. GzmB-triggered mammalian cell death is associated with activation of the BH3-only protein BID, disruption of the mitochondrial outer membrane, cytochrome c release, and caspase-9 activation. GzmB can directly process some caspase substrates, such as poly-(ADP-ribose)-polymerase (PARP), DNA-dependent protein kinase

(DNA-PK), ICAD, and lamin B [26]. GzmA and GzmB induce cell death not only in eukaryotic cells but also in intracellular bacteria, such as Escherichia coli, Listeria monocytogenes and

Mycobacteria tuberculosis, by disrupting their central metabolism and protein synthesis [27].

GzmB kills intracellular protozoan parasites (e.g, Trypanosoma cruzi, Toxoplasma gondii and

Leishmania major) by generating superoxide anions and inactivating oxidative defense enzymes

[28]. However, whether cytotoxic cells can use Gzms to kill intraerythrocytic P. falciparum remains unclear. P. vivax-infected reticulocytes are susceptible to GNLY-mediated membrane disruption because of cholesterol depletion from the host cell membrane. As a consequence, Gzms enter the infected cells, kill the parasite and lyse host reticulocytes [23]. P. falciparum-infected RBCs may also be susceptible to GNLY. However, whether GzmB can initiate programmed cell death in P. falciparum and what pathways are targeted have not been studied.

Figure 2: granzymes and their targets [29] Granzymes induce target cell death through diverse, non-redundant pathways that function in a complementary manner. GzmA promotes cell death by targeting mitochondria and the nucleus. Gzm B cleaves target cell proteins after aspartate residues like caspases. GzmB is known to induce caspase-dependent apoptosis by directly cleaving caspase 3 and 7 or indirectly by cleaving BID, disrupting the mitochondrial outer membrane, which leads to cytochrome c release and the assembly of the apoptosome. GzmB also targets the nucleus and induces DNA damage by activating CAD.

1.4.2 NK cells NK cells, which typically constitute ~10% of peripheral blood mononuclear cells (PBMCs), respond to P. falciparum. The cytotoxicity of NK cells is induced by cytokine stimulation, antibody/CD16 engagement, and the loss of inhibitory signaling due to the missing or mismatch of

MHC class I molecules. NK cells express killer cell immunoglobulin-like receptors (KIRs) that bind to MHC class I molecules [30], natural cytotoxicity receptors (NCRs, e.g. NKp30, NKp46) that recognized pathogen associated ligands, and killer lectin-like receptors (KLRs) such as

NKG2A and NKG2D that interact with HLA-E and MICA/B, respectively [31]. To respond to P. falciparum, NK cells require activating cytokines such as IL15, IL2, IL12/18 produced by dendritic cells, CD4 T cells and macrophages, respectively [32]. NK cells produce IFN-γ that stimulates macrophages within 6 hours of co-culture with P. falciparum-infected erythrocytes [33]. The IFN-γ response is heterogeneous between individuals [34]. A subset of NK cells that express high levels of CD94/NKG2A respond to infected erythrocytes [35] and form stable conjugates with infected erythrocytes in vitro [36]. NK cells lyse P. falciparum-infected RBCs and inhibit parasite growth via ADCC [37]. IgG from adults living in a malaria-endemic region, such as IgG to PfEMP1, promote NK cell cytotoxicity [37]. The capacities to secrete inflammatory cytokines and to lyse infected erythrocytes make NK cells good defenders against intraerythrocytic P. falciparum.

However, the cytotoxicity of NK cells depends on the presence of iRBC-specific opsonizing antibodies, which are only produced in late stage infection. The observation that freshly isolated human NK cells form conjugates with iRBCs suggests that NK cells might directly recognize some infection-related molecular patterns. Yet what receptors and ligands are involved in recognition are not well characterized and the heterogeneity in NK cell response to malaria remains largely unexplained.

1.4.2 γδ T cells

Unlike NK cells that kill iRBCs mainly by the engagement of CD16 with opsonizing antibodies, γδ

T cells possess TCRs that enable them to function independently of antibodies. γδ T cells are generated mainly in the thymus during fetal development, when common progenitor lymphocytes successfully rearrange the γδ chains via RAG-mediated V(D)J recombination at the double negative stage. The expression of αβ and γδ receptors on T cell progenitors is mutually exclusive: γ and β genes are encoded in different loci, while the δ genes, which are encoded in the cluster of α loci, is deleted when an α chain are rearranged. The diversity of the γδ TCR repertoire is determined before birth, and the majority of circulating γδ T cells share the same germline γ chain rearrangement [38].

γδ T cells are mainly found in tissues, such as the epidermis, dermis, intestine, lung and uterus, where they provide protection by sensing pathogen or danger associated molecular patterns

(PAMPs or DAMPs) in a manner reminiscent of innate immune cells. In this study, we focus specifically on the γδ T cells in the blood, most of which express the Vγ9Vδ2 TCR and comprise

~1-10% of circulating T lymphocytes. Vγ9Vδ2 T cells recognize non-peptide antigens in an MHC- independent manner. Vγ9Vδ2 T cells undergo rapid clonal expansion upon activation by prenyl pyrophosphate metabolites termed phosphoantigens (PAgs) from the eukaryotic mevalonate pathway (e.g. Isopentenyl pyrophosphate IPP) or prokaryotic isoprenoid synthetic pathway (e.g.

(E)-4-hydroxy-3-methyl-but-2-enyl pyrophosphate HMBPP) (Figure 3.a) that are “presented” by butyrophilin 3A1 (BTN3A1), which are type-I membrane proteins in the B7 superfamily with two

Ig-like extracellular domains [39]. Both isoprenoid synthesis pathways produce IPP, however, only the MEP/DOX-P pathway produces HMBPP, which is over 100 fold more potent than IPP at activating Vγ9Vδ2 T cells. The different sensitivity of Vγ9Vδ2 T cells to endogenous and exogenous antigens enables them to react to infection. The cytosolic domain of BTN3A1 which

10 possesses a positively charged pocket, binds a range of negatively charged intracellular PAgs to cause a conformational change in the extracellular portion of BTN3A1 [40], which is then recognized by the γδ TCR. Despite the difference of TCR, γδ T cells function like conventional T cells by secreting growth factors and cytokines required for stimulating innate cells, helping B cells develop, and killing target cells. It is controversial whether γδ T cells or NK cells are the predominant sources of IFN-γ in blood stage malaria patients. Vγ9Vδ2 T cells show polyclonal expansion during primary P. falciparum and P. vivax infection, reaching frequencies as high as

30% of circulating T cells [41]. These cells sense P. falciparum-derived PAgs, expand and express

GzmB and pro-inflammatory cytokines, such as IFNγ and TNFα [42]. Vγ9Vδ2 T cells proliferate after coculture with infected RBCs, schizonts and iRBCs culture supernatants [43]. Vγ9Vδ2 T cells can also be activated by PAgs secreted by gametocytes [44]. Other reports also suggest that the activation of Vγ9Vδ2 T cells can be stimulated by P. falciparum-derived phosphoantigens and degranulate without physically contacting iRBCs [45]. Vγ9Vδ2 T cells also recognize ligands such as F1-ATPase, endothelial protein C receptor, and toll-like receptors (TLRs). NKG2D on Vγ9Vδ2

T cells is an important co-stimulator that binds to MHC class I related proteins (MIC) A/B and the

UL16-binding proteins (ULBP1-4), which are often expressed on infected cells [46]. Vγ9Vδ2 T cells express CD16 and can recognize opsonized targets. However, Vγ9Vδ2 T cell-mediated

ADCC hasn’t been reported in malaria. Expanded Vγ9Vδ2 T cells increase expression of GNLY

[47] and inhibit P. falciparum growth in vitro [42]. Both iRBCs with late schizonts and merozoites cause Vγ9Vδ2 T cell degranulation [48]. Released GNLY might be important for intraerythrocytic

P. falciparum clearance [49].

Figure 3: phosphoantigen biosynthesis and γδ T cell functions [50] a. The mevalonate pathway (left) is found in archaebacteria, eukaryotes, and the cytoplasm of plants. The MEP/DOX-P (right) pathway is found in most eubacteria, apicomplexan protozoa and chloroplasts. Both pathways produce IPP, however, HMBPP is only produced via the MEP/DOX-P pathway. HMBPP is 100 times more potent than IPP at activating Vγ9Vδ2 T cells. b. After exposure to iRBCs, Vγ9Vδ2 T cells are activated by parasite-derived PAgs. Following activation, Vγ9Vδ2 T cells influence myeloid cell differentiation and activation by secreting cytokines and growth factors. After repeated parasite exposure, Vγ9Vδ2 T cells decrease production of proinflammatory cytokines and increase expression of CD16 and immunoregulatory molecules, such as Tim3. The expression of CD16 enables alternative functions, such as ADCC or antigen-dependent cellular phagocytosis (ADCP).

1.5 Research Goal

Our laboratory previously showed that T lymphocytes recognize P. vivax-infected reticulocytes to lyse the host cell and directly kill the parasite, preventing it from infecting new cells and spreading the infection. Lysis of the reticulocyte and parasite killing depended on HLA class I expression induced in the infected cell by malaria infection. Host cell lysis and P. vivax killing are mediated by granulysin and a granzyme. Because P. falciparum infects mature red blood cells that cannot induce HLA expression, CD8 T cells do not recognize blood stage malaria. However, blood

Vγ9Vδ2 T cells cells are activated in P. falciparum patients and recognize a PAg produced during malaria isoprenoid biosynthesis presented by BTN3A1 independently of HLA. Based on these data,

I hypothesize that Vγ9Vδ2 T cells recognize and lyse P. falciparum iRBC and kill the parasite in a

GNLY and GzmB dependent manner. The goal of my thesis is to test this hypothesis and begin to understand how the parasite is killed.

Chapter 2: Results and Methods

2.1 Results

2.1.1 Vγ9Vδ2 T cells target P. falciparum infected RBCs in a contact dependent manner

Killer cells release cytotoxic effectors when they recognize target cells. To test whether

Vγ9Vδ2 T cells degranulate in response to P. falciparum infection, Dr. Junqueira and I co-cultured cytokine stimulated Vγ9Vδ2 T cells with infected or healthy RBCs at an effector: target (E:T) ratio of 1:10. Uninfected RBCs did not trigger Vγ9Vδ2 T cell activation, but trophozoite P. falciparum- infected RBCs significantly stimulated degranulation (Figure 4.a). We examined whether Vγ9Vδ2

T cells can kill infected host cells by co-culturing Vγ9Vδ2 T cells with CFSE-labeled iRBCs or uRBCs. The CFSE fluorescence of iRBCs, but not uRBCs, was reduced after co-culture, indicating iRBCs were lysed. The lysis of iRBCs increased with increased E:T ratios (Figure 4.b). The specific lysis of P-falciparum infected RBCs is likely because Vγ9Vδ2 T cells recognize certain parasite- derived molecular patterns on infected host RBCs. To test if cell-cell contact is required for

Vγ9Vδ2 T cell-mediated killing, we did the co-culture experiment in a transwell plate. Vγ9Vδ2 T cells did not degranulate nor lyse iRBCs when physical contact was blocked (Figure 4.c, d).

2.1.2 Granzyme B, together with granulysin, leads to the death of P. falciparum

Our previous study showed that CD8+ T cells use GNLY to permeabilize the reticulocyte membrane and deliver Gzms into intracellular P. vivax. I hypothesized that GNLY might also act against intraerythrocytic P. falciparum. To assess the contribution of GNLY and GzmB to the death of P. falciparum, Dr. Junqueira and I treated trophozoite P. falciparum-infected RBCs with different combinations of purified GzmB, GNLY and PFN. The lysis of host RBCs was measured by LDH release. GNLY caused lysis of infected RBCs, and addition of GzmB led to significantly higher LDH release (Figure 5).

Figure 4: Vγ9Vδ2 T cells degranulate and kill upon co-cultured with iRBCs in a contact- dependent manner a. Cytokine stimulated Vγ9Vδ2 T cells were co-cultured with uRBCs or iRBCs in the presence of anti-CD107a for 4 hours. The percentage of degranulating cells is measured by the externalization of CD107a on δ2+ CD3+ cells. b. The number of lysed iRBCs, characterized by the loss of iRBC CFSE intensity, is proportional to the number of effector cells that were added. Cytokine stimulated Vγ9Vδ2 T cells were placed in the lower well of the bottom plate, while uRBCs or iRBCs were added in the top plate. a. Degranulation is measured as above. b. The cytotoxicity of Vγ9Vδ2 T cells was measured by the release of LDH from uRBCs/iRBCs.

Figure 5: purified GNLY and GzmB kill intraerythrocytic P. falciparum Sublytic doess of GNLY and PFN were titrated with healthy RBCs. GzmB (500 nM) and sub-lytic dose of GNLY (200 nM) / PFN (1/1000 dilution) were added to trophozoite P. falciparum-infected RBC culture. The lysis of iRBC was measured by the release of LDH in the supernatant after 4 hours incubation with purified proteins. Percent LDH release was calculated by subtracting background signal (medium blank) and normalized to maximum release (0.02% Triton).

2.1.3 Granzyme B Targets Various Proteins in P. falciparum

To study how GzmB kills the parasite, I treated P. falciparum lysates with 200 nM GzmB or lysis buffer and applied proteomics to find the substrates of GzmB in P. falciparum using the

Terminal Amine Isotopic Labeling of Substrate (TAILS) method. In brief, protein samples from

GzmB treated or untreated parasites were labeled with different isobaric TMT labels, trypsinized and analyzed by mass spectrometry. Peptides with one mass-reporter and with their N-terminal downstream of an Asp residue were considered GzmB-digested fragments (Figure 6). All of the predicted GzmB cleavage sites loosely fit the GzmB putative motif of IEPD. Enolase is required for glycolysis in intraerythrocytic parasites and invasion by merozoites and sporozoites. Purine nucleoside phosphorylase is involved in the purine salvage pathway that is essential for P. falciparum survival. Falcipain 2, a papain family cysteine protease, and Plasmepsin IV, an aspartic protease, are required for hemoglobin hydrolysis in intraerythrocytic parasites food vacuole.

Uracil-DNA glycosylase is a specific DNA glycosylase for uracil cleavage in base excision repair and is involved in preventing G:C to A:T transition mutation. Mitogen-activated protein kinase 1 may coordinate with MAP kinase 2 and may be needed to the complete the P. falciparum asexual cycle. PTEX88 likely plays a specific role in mediating parasite sequestration and trafficking of exported proteins. GzmB also targets subunits, such as histone, tRNA (m1A) methyltransferase complex, and ribosome, of complexes that are essential for maintaining normal cellular functions.

I also utilized 2D electrophoresis and MS/MS to identify putative GzmB substrates. I was able to resolve 454 spots, and identify 21 spots that disappeared post GzmB digestion (Figure 7).

The identification of those spots is still pending.

Gene$! Name! Cleavage$Site$!Loca2on$! Func2on$! VEVD! cytoplasm,!mitochondria,!nucleus,!! ENO! Enolase! glycolysis!! IVGD!! cell!membrane,!cytoskeletal! PF3D7_0513300! Purine!nucleoside!phosphorylase! VVCD! cytoplasm,!mitochondria!! nucleoside!metabolic!process!

PF3D7_1115700! Cysteine!proteinase!falcipain!2a! IETD! food!vacuole,!lysosome! hemoglobin!proteolysis!

PF3D7_1407800! Plasmepsin!IV! IESD! food!vacuole! hemoglobin!proteolysis!

PF3D7_1333200! UbiquiOnPacOvaOng!enzyme! LTDD! Cytoplasm,!apicoplast! ubiquiOnPprotein!transferase!

PF3D7_1415000! UracilPDNA!glycosylase! IEND! mitochondrion,!nucleus! DNA!hydrolysis!

PF3D7_1431500! MitogenPacOvated!protein!kinase! ILTD! cytoplasm,!nucleus! MAP!kinase,!protein!phosphorylaOon!

PF3D7_1105600! Translocon!component!PTEX88! IDVD! vacuole!PTEX!! protein!transportaOon! chromosome!organizaOon,! PF3D7_1105000! Histone!H4! ILRD! nucleus! nucleosome!assembly!

Ribosome!maturaOon!protein! PF3D7_1410900! LNID! cytoplasm!! ribosome!biogenesis! SBDS,!putaOve!

WD!repeatPcontaining!protein,! PF3D7_1410300! VNTD! 90S!preribosome! SSUPrRNA!maturaOon!! putaOve!

tRNA!methyltransferase!subunit! PF3D7_0912800! IYDD! nucleus! tRNA!methylaOon! TRM6,!putaOve!

Figure 6: candidate substrates of GzmB in P. falciparum lysate GzmB treated and untreated samples were tagged with TMT-duplex labels (TAILS). Identified amino acid sequences of each detected peptide were compared/aligned to the proteome database of P. falciparum 3D7 on UniProt. Standard filters with Rank=1, PPM -5~5, ΔCorr ≥ 0.1, XCorr ≥1 were applied to exclude false positive reads. Proteins were selected as GzmB substrates when the identified peptides were predicted to follow an Asp residue and were labeled with one mass reporter signal. Information on localization and functionality of each candidate protein were derived from UniProt and KEGG. The order of listed candidates was determined by the score of sequence alignment to existing database and the order does not represent the efficiency or specificity of GzmB cleavage.

18 pH3$ pH10$ reduced$MW$$

Figure 7: 2D-gel electrophoresis identification of substrates in P. falciparum lysate Protein samples from the untreated or GzmB-digested parasite lysates were separated by isoelectric focusing (PH 3-10 from left to right) in the first dimension and by gradient SDS-PAGE in the second dimension. Spots present in the untreated sample, but missing in the GzmB treated sample were considered candidate substrates, marked with red squares. The candidates were excised, in gel digested, sent to MS/MS and analysis is pending.

2.1.4 GzmB disrupts glycolysis P. falciparum

To validate the putative GzmB substrates, I expressed the protein candidates in vitro and tested for their cleavage by GzmB. I started with enolase, an enzyme required for producing pyruvate from D-glyceraldehyde 3-phosphate by converting 2-phospho-D-glycerate (2-PG) into phosphoenol pyruvate (PEP) during glycolysis and gluconeogenesis. To test if Pfeno is indeed cleaved and whether glycolysis is disrupted by GzmB, I treated recombinant Pfeno (rPfeno) with

GzmB (Figure 8.b). GzmB cleaved rPfeno and produced fragments of sizes that matched my predicted cleavage site (Figure 8.a, c). To assess if GzmB digestion abolishes the enzymatic function of rPfeno, I measured the production of PEP from 2-PG. Because Pfeno requires Mg2+ as a co-factor, I measured the catalytic activity of Pfeno with various concentrations of Mg2+. I found that GzmB treatment reduced the catalytic efficiency of rPfeno at various Mg2+ concentrations

(Figure 8.d).

2.1.5 GzmB disrupts the purine synthetic pathway of P. falciparum

I next assessed cleavage of P. falciparum purine nucleoside phosphorylase (PfPNP), which is found in the cytoplasm and mitochondria. Plasmodium species are purine auxotrophs and depend on purine salvage for growth and replication [51]. PfPNP catalyzes the conversion of inosine to hypoxanthine, the precursor of all purines (Figure 9.a). Defect in PfPNP significantly impacts the growth of P. falciparum. I expressed and treated recombinant PfPNP (Figure 9. b) with

GzmB. GzmB digested PfPNP and produced two fragments of the predicted sizes (Figure 9. c).

a" b"

c" d" kD" 250"

150" 100" 75"

Full"length"48.7kD" 50"

Fragment"1:"33kD" 37"

25" 20"

" 15" Fragment"2:"13kD" " 10" " Figure 8: GzmB cleaves P. falciparum enolase and disrupts glycolysis a. The amino acid sequence of Pfeno is shown at the left, with red tetrapeptide sequences representing the cleaving sites detected by TAILS-MS/MS. b. BL21 E. coli was transformed with plasmids pET-24a(+) containing bacterial codon-optimized rPfeno sequence (Codon Adaptation Index CAI=0.79, average GC content= 41.24%, designed according to GenScript OptimumGene algorithm). rPfeno was induced by 10 mM IPTG and purified with Ni-NTA resin. c. Purified rPfeno was digested with 200 nM GzmB for 15 minutes at 37oC. d. The enzyme activity of rPfeno was assessed by measuring the conversion of 2-PGA to PEP by measuring the increase in absorbance between 230 and 240 nm.

b kD5 c 2505 PNP# 1505 1005 MDNLLRHLKISKEQITPVVLVVGDPGRVDKI 755 KVVCDSYVDLAYNREYKSVECHYKGQKFLC 505 VSHGVGSAGCAVCFEELCQNGAKVIIRAGS 375 CGSLQPDLIKRGDICICNAAVREDRVSHLLIH Full5length5275kD5 GDFPAVGDFDVYDTLNKCAQELNVPVFNGI 255 SVSSDMYYPNKIIPSRLEDYSKANAAVVEME 205 Fragment51:5225kD5 LATLMVIGTLRKVKTGGILIVDGCPFKWDEG DFDNNLVPHQLENMIKIALGACAKLATKYA5 155

105 Fragment52:565kD5

Figure 9: GzmB disrupts the purine metabolic pathway [52] a. Purine bases and nucleosides are taken up by erythrocyte nucleoside transporters. P. falciparum PfADA and PfPNP enzymes salvage purines and recycle methylthioadenosine, a product of polyamine synthesis. Both purine bases and nucleosides are taken up by the parasite-encoded nucleoside transporter PfNT1. The purine salvage pathway includes PfADA-mediated deamination of adenosine to inosine, PfPNP-mediated conversion of inosine to hypoxanthine, and HGXPRT- mediated purine production. b. The sequence of PfPNP, the cleavage site identified by TAILS- MS/MS is highlighted. c. rPfPNP was induced by 1mM IPTG and purified with Ni-NTA affinity purification. rPfPNP was incubated with 200 nM GzmB for 15 min at 37oC.

2.1.6 GzmB targets the mitochondrion of P. falciparum

GzmB is known to cause rapid oxidative death in bacteria by disrupting electron transport and oxidative phosphorylation, generating superoxide anion and crippling bacterial oxidative defenses [53]. The P. falciparum mitochondrion is important for essential metabolism and pyrimidine biosynthesis. Therefore the disruption of the mitochondrion should be detrimental for parasite survival. Both Pfeno and PfPNP are localized to the parasite mitochondrion, suggesting

GzmB may localize to and target the mitochondrion. Since P. falciparum possesses an electron transportation chain that shares functional similarities with that in bacteria, I hypothesized that

GzmB might also disrupt the mitochondrion in P. falciparum. To test this hypothesis, I isolated and treated crude parasite mitochondria with GzmB and assessed mitochondrial protein cleavage by 2D gel electrophoresis. Spots that were present in the untreated sample but missing in GzmB- treated mitochondria were considered candidate substrates. Based on mass spectrometry, 10 of 13 spots were identified: enolase and phosphoglycerate mutase that are involved in glycolysis; ATP synthase subunit beta and ubiquinol-cytochrome-c reductase complex assembly factor 1 that make up electron transport chain complexes; hypoxanthine phosphoribosyltransferase which is involved in purine salvage pathway; elongation factor 1, eukaryotic initiation factor 2A and heat shock proteins that are important for translation and protein folding; and transportation-related proteins, such as vacuolar protein sorting-associated protein 4 and GTP-binding nuclear protein (Figure 10).

To examine if GzmB disrupts the P. falciparum mitochondrion, I also measured the parasite extracellular oxygen consumption rate (OCR) of mitochondria incubated or not with GNLY and

GzmB. GzmB alone reduced the OCR. The addition of GNLY further reduced the OCR in a dose dependent manner (Figure 11), suggesting that GNLY and GzmB disrupt mitochondrial activity.

Protein(( Loca,on(( Func,on(

Enolase(( (( cytoplasm,(mitochondria,(nucleus,(( glycolysis(( cell(membrane,(cytoskeletal( ATP(synthase(subunit(β(( mitochondria(( ATP(synthesis((

Phosphoglycerate(mutase(( cytoplasm( glycolysis(( ( Hypoxanthine( cytoplasm( hypoxanthine(metabolism,(purine( phosphoribosyltransferase(( salvage( Phosphoethanolamine(N@ Golgi(apparatus( phosphaBdylcholine(biosyntheBc( methyltransferase( process( Vacuolar(protein(sorBng@associated( cytoplasm( vacuole(organizaBon,(protein( protein(4( transportaBon( Heat(shock(protein(( cytoplasm,(nucleus(( protein(refolding(

ElongaBon(factor(1@alpha (( cytoplasm( translaBonal(elongaBon((

EukaryoBc(iniBaBon(factor(4A(( cytoplasm(( translaBon(iniBaBon((

GTP@binding(nuclear(protein( cytoplasm,(nucleus(( nucleocytoplasmic(transport(

Ubiquinol@cytochrome@c(reductase( mitochondria( mitochondrial(respiratory(chain( complex(assembly(factor(1,(putaBve( complex(III(assembly( Figure 10: 2D-gel electrophoresis identification of candidate substrates from crude mitochondrial fractions Protein samples from untreated group or GzmB digested group were separated by isoelectric focusing (PH 3-10 from left to right) in the first dimension and by gradient SDS-PAGE in the second dimension. Spots that were present in untreated mitochondria but missing in the GzmB treated sample were considered candidate substrates, marked with squares. The candidate spots were excised, in gel digested and sent to MS/MS for analysis. Results from mass spectrometry were checked by comparing the molecular weight and pI of the protein hits to the corresponding spots on the gel. Proteins with 2 to 5 unique peptides were selected.

GNLY%&%GzmB%disrupt%P.#falciparum#mitochondrion%

!!!!!!!!!GNLY!(nM)!!!!!!0!!!!!!!!!!!!!!!!!0!!!!!!!!!!!!!!!10!!!!!!!!!!!!!!!!20!!!!!!!!!!!!!!50!!!!!!!!!!!!!200! An2mycin!A! ! 500nM!GzmB! !

Figure 11: GNLY and GzmB treatment inhibit extracellular oxygen consumption ~1 million 0.01% saponin-freed middle trophozoites were cultured in RPMI 1640 with 500 nM GzmB and GNLY at various concentrations for up to 30 minutes. All conditions were assessed in triplicate. The extracellular oxygen consumption rate was measured by measuring the change of extracellular oxygen concentration. Antimycin A, a natural toxin that inhibits complex III was used as a positive control. All results were normalized to the value of untreated mitochondria, which was set as 100%.

GzmB, GNLY,

Endosome,,, VPS4' Cytosol, Mitochondria, Transla/on, Purine,salvage, ETC,,ATP,synthesis,complex, SBDS' PfHGXPRT' , F"type'ATPase' WDR' PfPNP' Endoplasmic,re/culum, eEF1A' eIF4A' Nucleus,, ' Glycolysis,, Bip' Glycerophospholipid,biogenesis, Protein,exporta/on,,Ran' HSP70' PGM' HSP90' Nucleo/de,modiﬁca/on,, PfPMT' ENO' PfUDG' Histone'H4' TRM6' Food,vacuole,, Hemoglobin,proteolysis, FP2' Plm'IV' Protein,transporta/on, PTEX88'

Figure 12: putative targeted GzmB substrates identified by proteomics

Summary of all putative GzmB substrates, their cellular location and pathways involved. Ran GTP-binding nuclear protein (Ran), Uracil-DNA glycosylase (UDG), histone H4, tRNA methyltransferase subunit TRM6 (TRM6), endoplasmic reticulum chaperon Bip, HSP70 and HSP90, ribosome maturation protein (SBDS), WD repeat-containing protein (WDR), elongation factor 1 alpha (eEF1A), eukaryotic initiation factor 4A (elF4A), hypoxanthine phosphoribosyltransferase (HGXPRT), purine nucleoside phosphorylase (PNP), F-type ATPase subunit beta, Phosphoethanolamine N-methyltransferase (PMT), phosphoglycerate mutase (PGM), enolase (ENO), cysteine proteinase falcipain 2a (FP2), plasmepsin IV (Plm IV), translocon component PTEX88 (PTEX88) and vacuolar protein sorting-associated protein 4 (VSP4).

2.2 Materials & Methods

δ2 T Cell Isolation, Culture & Stimulation:

δ2 T cells were purified from peripheral blood mononuclear cells (PBMCs) of healthy donors.

Blood samples collected within 6 hours were diluted with RPMI 1640 to ~35ml, applied to 10ml

Ficoll, and centrifuged at 300g for 35 minutes with no brake. PBMCs at the interface were collected and washed with RPMI 1640, 200g for 15 minutes, and cultured with complete RPMI medium (10% human serum, glutamine and antibiotics) overnight. δ2 T cells were sequentially labeled with PE anti-δ2 antibody (BioLegend) and anti-PE microbeads (Miltenyi Biotec,

Germany), and positively selected with MACS (Miltenyi Biotec, Germany). Isolated δ2 T cells were cultured with complete RPMI medium containing IL2 (200 IU/ml) and IL15 (25 ng/ml) for 6 days, with medium replacement every 2 days. Purity of cultured cells was assessed by FACS staining for δ2, CD3, GzmB, GzmA, GNLY and PFN.

Plasmodium falciparum Culture and Synchronization:

P. falciparum 3D7 parasites were cultured in O+ human RBCs in RPMI 1640 with 25 mM HEPES buffer, 3.60 g/L glucose, 50 mg/L hypoxanthine, 21% sodium bicarbonate, 5% Albumex-II, 5%

o CO2, 5% O2 and 90% N2 at 37 C. The culture medium was changed every day to reach 2% hematocrit and parasitemia was maintained at 2-5%. Schizonts were enriched by 60% Percoll gradient, washed with complete medium and added to fresh RBC culture to reach a final 2% hematocrit. For synchronization, RBCs were harvested 6-8 hours post infection, incubated with 5% sorbitol for 15 mins at room temperature, washed and plated on a new dish. Highly pure trophozoites (30 hrs post infection) were enriched by MACS and used for all co-culture assays.

Cytotoxicity assay:

(1) LDH released from infected RBCs was measured by CytoTox 96 NonRadioactive Cytotoxicity

Assay (Promega). Culture supernatants were collected from 96 well plates at 400g for 5 minutes.

The amount of LDH in supernatants was assessed by measuring the increase of absorbance at 490 nm. (2) Flow cytometry-based analyses were used to measure the cytotoxicity of effector cells.

Infected/uninfected red blood cells were labeled with 5 uM CFSE (Sigma) and co-cultured with

Vγ9Vδ2 T cells at indicated E:T ratios. After 4 h co-culture, RBCs were collected, stained with

CD235a and CD3, washed and analyzed by flow cytometry. The percent loss of CFSE signal was normalized to the count beads added to each reaction tube.

Bacterial Transformation & Induced Protein Expression:

Bacterial codon-optimized plasmids encoding P. falciparum candidate granzyme B substrates

(Enolase, PNP etc.) were designed and ordered from GeneScript. Competent BL21 bacteria were mixed with plasmids, and allowed to sit on ice for 30 minutes. The transformation step was achieved at 42 oC for 90 seconds. The cooled mix was then allowed to grow in 1 mL SOC medium

(Sigma-Aldrich) for 1 hour at 37 oC with constant shaking, followed by overnight culture on plates with selecting antibiotics at 37 oC. A colony was picked and cultured overnight with antibiotic- containing Luria-Bertani broth (LB) at 37 oC. The culture was diluted into antibiotic-containing

LB and allowed to expand to an OD of 0.6. 10 mM IPTG was added to induce protein expression at 37 oC for up to 4 hours. Bacteria were collected by centrifugation at 16,000g for 3 minutes.

Recombinant Protein Purification:

IPTG treated cells were washed with ice cold STE buffer (150 mM NaCl, 1 mM EDTA, 10 mM

Tris-HCl; pH 8) and incubated in 800 µl STE buffer containing 100 ug/ml of lysozyme (Sigma-

Aldrich) on ice for 15 minutes. Bacteria were lysed by the addition of DTT and N-laurylsarcosine to a final concentration of 5 mM and 1.5% (w/v), respectively. Samples were sonicated (3 cycles, 5 minutes each) and the lysates were centrifuged at 16,000g for 10 minutes. Supernatants were incubated with Ni-NTA Superflow beads (Qiagen) at room temperature, and mixed for 40 minutes.

Beads were washed with 6 column volumes of 25 mM Imidazole and His-tagged proteins were

28 eluted with 300 mM Imidazole. Eluates were concentrated using 10K Amicon Ultracentrifugal

Filter units (Millipore).

P. falciparum Protein Purification:

Infected RBCs were isolated from P. falciparum 3D7 cultures with magnetic columns and lysed with lysis buffer (0.01% Saponin, 120 mM KCl, 20 mM NaCl, 20 mM glucose, 6 mM HEPES, 6

o mM MOPS, 1 mM MgCl2, 0.1 mM EDTA, pH 7.5) for 5 minutes at 4 C. Lysed iRBCs were washed with lysis buffer several times at 800g for 6 minutes. Pelleted parasites were then lysed using non-ionic lysis buffer (5% Triton X-100, 50 mM Tris-HCl, 150 mM NaCl; pH 7.5) and homogenized by passage through 28G and 30G needles. Hemozoins were removed by centrifugation at 20,000g, 4 oC for 10 minutes. Hemoglobin was removed by incubating the cleared protein solution with anti-hemoglobin resin (HemogloBind, Biotech Support Group) at 4 oC for 5 minutes. Cleared supernatants containing P. falciparum proteins were used for enzyme digestion assay.

P. falciparum Crude Mitochondrial Fraction Isolation:

P. falciparum were freed from infected RBCs using saponin and then resuspended in 20 mL mitochondria suspension buffer (225 mM mannitol, 75 mM sucrose, 0.1 mM EDTA, 3 mM Tris–

HCl; pH 7.4) and placed inside a pre-cooled nitrogen bomb. Nitrogen cavitation was performed at

12,000 psi for 20 minutes, followed by slow drop-wise release of solution containing subcellular fragments of disrupted parasites. Intact parasites were removed by centrifugation at 800g for 6 minutes and the crude mitochondrial fraction was collected by centrifuging the supernatant at

23,000g, 4 oC for 20 minutes.

Enzyme Digestion:

Protein samples were prepared by methanol/chloroform precipitation, followed by incubation in enzyme buffer (50 mM NaCl, 10 mM TrisHCl; pH 7.4) w/ or w/o 200 nM granzyme B at 37 oC for

15 to 20 minutes. The crude mitochondrial fraction was incubated with mitochondria resuspension buffer w/ or w/o 200 nM granzyme B at 37 oC for 20 minutes, and the pellets were collected by centrifugation at 21,000g for 15 minutes and lysed with RIPA buffer.

Protein Precipitation and Lyophilization:

(1) Methanol/chloroform precipitation was used to remove detergents in protein samples: protein samples were mixed with 4 sample volumes ice cold methanol, 1 sample volume chloroform and 3 sample volumes H2O, and centrifuged at 14,000g for 2 minutes. The white pellets at the interface were washed with 4 sample volumes methanol and collected by centrifugation at 20,000g for 5 minutes. The pellets were allowed to dry for 5 minutes and stored at -80 oC for further use. (2)

Acetone precipitation was used to remove excess TAILS reagents: protein samples were mixed with acetone (-20 oC) and allowed to precipitate overnight at -20oC. Precipitants were collected by centrifugation at 20,000g for 5 minutes. (3) Freeze-drying lyophilization was used to concentrate post-TAILS samples: protein samples were pre-frozen at -80oC and sublimation was carried out at

0.0141 mBar, -60 oC until no frozen material was left. The concentrated samples were dialyzed and desalted.

2D Gel Electrophoresis:

Protein samples were prepared using ReadyPrep 2D Cleanup Kit (Biorad) and actively rehydrated

(9 M Urea, 3% CHAPS. 30 mM DTT, 1X Bio-Lytes, constant 50V, 14h) into ReadyStrip IPG

Strips (7 cm NL3-10, Biorad). First dimension isoelectric focusing (IEF) was carried out in Biorad

Protean IEF Cells in three steps: (1) conditioning (250 V, 15 minutes), (2) rapid voltage ramping

(4,000 V, 2 hours) and (3) final focusing (4,000 V, 20,000 vhours). The strips were kept at -80 oC for further use, or equilibrated with SDS equilibration buffer (6 M Urea, 375 mM Tris, 2% SDS,

20% glycerol; pH 8.8) containing (1) 1,4-Dithiothreitol (DTT, w/v 2%, Sigma-Aldrich) and (2)

Indole-3-acetic acid sodium salt (IAA, w/v 2.5%, Sigma-Aldrich) for 15 minutes each. The

30 equilibrated strips were then rinsed with SDS-PAGE running buffer, laid on top of Gel IEF precast gels (7 cm strip, Biorad) or 12% SDS-PAGE gel (17 cm strip) with Precision Plus Protein

Unstained Protein Standard Plugins, and sealed with 1% low-melt agarose containing 0.1%

Bromophenol Blue. Second dimension separation was performed at 100 V for 2 h or 25 mA/per gel overnight until the dye front reached the bottom line.

Coomassie Blue & Silver Staining:

SDS-PAGE gels were fixed with 40% ethanol and 10% acetic acid for 40 minutes and stained with

Coomassie R250 solution (0.1% Coomassie Blue R250, 10% acetic acid, 50% methanol) for 1 hour. Gels were destained with 50% methanol, 10% acetic acid. Silver staining was performed with SilverQuest kit (Invitrogen) at room temperature according to the product instructions.

Terminal Amine Isotopic Labeling of Substrates (TAILS):

~100 µg protein samples from cell lysates or crude mitochondrial fractions of P. falciparum were precipitated with methanol/chloroform and dissolved in 100 µl of 50 mM Triethylammonium bicarbonate (TEAB, Sigma-Aldrich) buffer. Protein reduction and alkylation were performed with

5 µl of 200 mM Tris (2-carboxyethyl) phosphine hydrochloride (TCEP, Sigma-Aldrich), 55 oC for

1 h and 5 µl of 375 mM 3-Indolacetic acid (IAA, Sigma-Aldrich), room temperature for 30 minutes. Protein samples were labeled with 24 µl of TMT reagents (0.8 mg TMT126/127 in anhydrous DMSO, ThermoFisher Scientific) at room temperature for 2 h and quenched with 16 µl

5% hydroxylamine (Sigma-Aldrich). TMT labeled samples were pooled, washed with acetone and trypsinized with 5 µg of Sequencing Grade Modified Trypsin (Promega) overnight at 37 oC. TMT- labeled peptides were incubated with 400 µl of immobilized anti-TMT antibody resin (Thermo

Scientific), washed with 1X TBS and pure water and eluted with 4 column volumes of TMT elution buffer (ThermoFisher Scientific). Eluates were concentrated by lyophilization.

Mass Spectrometry:

Spots that disappeared after GzmB digestion were excised from the gel, in-gel trypsin-digested and analyzed on a nanoscale reverse-phase HPLC by electrospray ionization and LTQ linear ion-trap mass spectrometry at the Taplin Biological Mass Spectrometry Facility, Harvard Medical School.

Results from mass spectrometry were checked by comparing the molecular weight and pI of the protein hits to the corresponding spots on the gel. Proteins with 2 to 5 unique peptides were selected for analysis. TAILS samples were compared to the proteome database of P. falciparum

3D7 and gated by Rank=1, PPM -5~5, ΔCorr ≥ 0.1, XCorr ≥1.

Enolase enzyme activity assay

The activity of Pfeno was measured by the conversion of 2-PG (Sigma) to PEP. 25 µL of assay

mixture containing 2 mM 2-PGA, 5 µg Pfeno, 1- 20 mM MgCl2, 50 mM Tris/HCl, pH 7.5, was used. The production of PEP was measured by the increase in absorbance at 230-240 nm using

Nanodrop.

Extracellular oxygen consumption

P. falciparum mitochondrion activity was measured by Extracllular O2 Consumption Assay

(Abcam). Typically 10 µl of reagent was added to 150 µl RPMI medium containing ~1 million saponin-freed parasites, and the wells were immediate sealed with high sensitivity mineral oil. The fluorescence change was monitored with dual-read time-resolved fluorescence, top read, Ex 380 ±

20 nm, Em 645 ± 15 nm, with a time interval of 2 minutes for 1 hour using a BioTek Synergy 2 plate reader.

Chapter Three: Discussion

Previous data from our lab show that killer cells use GNLY to deliver Gzms into bacteria and P. vivax [26, 30]. Here I show that Vγ9Vδ2 T cells respond to P. falciparum infected RBCs by releasing their cytotoxic mediators (Figure 4. a) and lysing infected RBCs (Figure 4. b). Cytokine primed Vγ9Vδ2 T cells degranulate in response to P. falciparum-infected, but not uninfected,

RBCs, indicating certain parasite-derived molecular patterns may be recognized by Vγ9Vδ2 T cells. The fact that both Vγ9Vδ2 T cell degranulation and RBC lysis can be prevented when co- culture was performed in a transwell plate (Figure 4. c & d) suggests that cell-cell engagement is required for the activation of Vγ9Vδ2 T cells. To study what cytotoxic mediators from Vγ9Vδ2 T cells contribute to the lysis of infected host cells, I treated trophozoite P. falciparum-infected

RBCs with different combinations of purified Pfn, GNLY, and GzmB. By measuring LDH release from host RBCs, I show that the addition of sublytic GNLY to uninfected RBCs can lyse infected

RBCs, and presumably disrupt the membrane of intraerythrocytic parasites. The combination of

GzmB with GNLY caused significantly higher LDH release compared to all other groups (Figure

5), suggesting that GNLY facilitates the delivery of GzmB into intraerythrocytic parasites. Dr.

Junqueira has reported that P. vivax-infected reticulocytes with reduced surface cholesterol are targeted by GNLY [23]. Host cell cholesterol depletion may also occur in P. falciparum infection to explain the susceptibility of infected RBCs to GNLY. Pfn seems to also disrupt the membrane of infected RBCs. To test which pore-forming protein contribute more to the delivery of GzmB into the parasite, I plan to use fluorescent GzmB and compare the entry of GzmB with GNLY or

Pfn.

GzmB is a selective protease that cleaves after Asp residues within preferred tetrapeptide sequences of IEPD. The binding pocket of the GzmB is not very deep and the substrate specificity is determined by extended binding of substrates rather than the motif alone [54]. It is unclear if any

33 tetrapeptide-based algorithm could provide in-silico predictions of GzmB substrates with high confidence. To understand how Vγ9Vδ2 T cells kill intraerythrocytic parasites, I used proteomics to identify candidate GzmB substrates in P. falciparum. Previous work from our lab showed that

GzmB can initiate oxidative cell death and disrupt central metabolism of bacteria. My work with

P. falciparum shows that GzmB also disrupts parasites metabolism, especially glycolysis and the purine-salvage pathway. My preliminary proteomic analysis suggests that GzmB also targets proteins in diverse cellular locations, and disrupts cellular functions important for parasite survival, including nucleotide modification, glycerophospholipid biogenesis, protein synthesis, degradation and transportation (Figure 6). TAILS-MS/MS is high throughput analysis under ideal conditions.

However, I did not obtain optimal result because residual hemoglobin interfered with labeling.

Importantly, all substrates identified by this method have cleavage sites that match the GzmB preference for Asp at the P1 position and a neutral amino acid (L, I or V) at the P4 position, suggesting they are bonafide substrates. Some of these GzmB-disrupted pathways could be considered for new anti-malarial drug targets.

Prediction of GzmB substrates was made based on mass-spectrometry detection of unique peptides and sequence alignment to the P. falciparum database. To test if the candidates are indeed digested by GzmB, I performed in vitro cleavage assays. Because asexual proliferation of P. falciparum depends highly on glycolysis, I first tested whether enolase is a GzmB substrate. I showed that recombinant P. falciparum enolase was cleaved by GzmB to generate peptides of sizes close to my prediction (Figure 8. a & c). Unlike host cells that have 3 subunits of the enolase monomer and 3 isoenzymes, P. falciparum only expresses one enolase isoform that forms a homodimer. Incubation of Pfeno with GzmB significantly inhibits its enzymatic activity, which was measured by the conversion of 2-PG to PEP (Figure 8. d). Another putative GzmB substrate, phosphoglycerate mutase (PGM) that catalyzes the conversion of 2-PG from 3-PG, is also worth

34 validating. P. falciparum expresses few enzymes for glycolysis. Instead of a human-homologous phosphofructokinase (PFK), they use a plant-like 6-phospho-1-fructokinase Disrupting Pfeno and

PfPGM should strongly inhibit glycolysis in intraerythrocytic P. falciparum and contribute to the death of parasites. No one has successfully generated a Pfeno knock out strain, presumably because enolase is essential for parasite survival. To investigate if GzmB-mediated cleavage of

Pfeno inhibit glycolysis, I plan to measure the glucose uptake rate as well as the extracellular acidification rate.

I next showed that GzmB cleaved P. falciparum purine nucleoside phosphorylase (Figure

9), a target of quinine and mefloquine [55]. P. falciparum cannot produce purine rings de novo and relies on PfPNP, PfADA, and PfHGXPRT that act non-redundantly for salvaging host purines.

PfPNP is required for the second step of the pathway, and it is also important for the homeostasis of polyamine synthesis. Disruption of PfPNP activity, such as treatment with transition state inhibitors (Immucillins), causes the purine starvation-induced death of P. falciparum both in vitro and in vivo [56]. Cleavage of PfPNP by GzmB should abolish its function. PfPNP shares functional similarity with host PNP, and a defect in PfPNP causes reduced growth rate rather than parasite death, presumably because P. falciparum can utilize intermediates produced by host purine salvage pathways.

GzmB is known to target eukaryotic mitochondria. The Plasmodium mitochondrion has the smallest eukaryotic mtDNA and highly fragmented rRNA genes. Because the parasite mitochondrion differs from host mitochondrion, it is a target of various antimalarial drugs [57].

Current antimalarial drug development focuses on inhibitors of the mitochondrial electron transport chain (ETC), including ETC complexes III, IV, and V that generate electron gradients on the mitochondrial inner membrane, and proteins in the pyrimidine biosynthesis pathway of P. falciparum. Understanding how GzmB disrupts the P. falciparum mitochondrion may point to

35 novel drug targets. Because of the difficulty obtaining pure mitochondria, I identified putative

GzmB substrates in the “crude mitochondrial fraction”, which contains mitochondrion and other organelles such as the food vacuole and apicoplast (Figure 10). I also observed the crude mitochondrial fraction were lyzed after the incubation with GzmB, which suggests that GzmB may target proteins required for mitochondrial osmotic balance. To further investigate how GzmB disrupt parasite mitochondria, I plan to isolate highly pure mitochondria with gradient centrifugation and check if GzmB treatment causes morphological changes using electron microscopy.

Treatment with GNLY and GzmB inhibited extracellular oxygen consumption (Figure 11), indicating that GzmB may require GNLY to enter the parasite mitochondrion. Treatment with

GzmB alone also inhibited oxygen consumption, which is likely due to the membrane disruptive effect of saponin, which was used for releasing parasites from host RBCs. However, this result does not indicate if GNLY alone also disrupts mitochondrial function since GNLY can also disrupt mitochondrial membrane and proton gradient. To improve the assessment of GzmB effect on

OCR, I plan to repeat the experiment using digitonin, an alternative to GNLY that has neglectable impact on parasite mitochondria membrane integrity [58]. I also plan to assess the GzmB effect on mitochondria of early schizonts, which have higher OCR and relatively more permeable membranes compared to trophozoites. In this case, I will only treat the free schizonts with GzmB. I also plan to assess the overall glycolysis of GNLY and GzmB treated parasites by measuring the extracellular acidification rate (ECAR). By combining the OCR and ECAR, I can assess the metabolic changes of GNLY and GzmB treated P. falciparum.

Although I have validated two predicted GzmB candidates in vitro, the results might not reflect how GzmB works in vivo. It is unclear which cellular compartments GzmB can access and whether it can digest proteins inside different organelles. To answer these question, I plan to treat

36 the parasites with fluorescent GzmB to find the active location of GzmB by fluorescent microscopy. I will validate more predicted GzmB candidates at different cellular location, such as the ATP synthase subunit beta on mitochondria, phosphoethanolamine N-methyltransferase in cytosol, plasmepsin V in the food vacuole, elongation factor 1 alpha that is associated with ribosome. To determine whether GzmB cleaves putative substrates and disrupts pathways in vivo, I will assess the cleavage of proteins expressed in E.coli with western blotting. I will also treat P. falciparum that produces candidate proteins or non-cleavable mutants with GNLY and GzmB to verify that predicted targets are indeed cleaved, and assess whether the cleavage contributes to parasite death.

Unlike antimalarial drugs that generally target a single essential protein, GzmB seems to utilize a multipronged strategy to kill malaria parasites (Figure 12). Applying this proteomics approach to other malaria parasites, especially P. vivax, and comparing the profile of GzmB substrates and targeted pathways may identify common targets or targeted pathways. A promising way to develop antimalarial drugs to intraerythrocytic P. falciparum would be to mimic granzymes by designing proteases that could disrupt multiple metabolic and biosynthetic pathways. A fusion protein that links GzmB to merozoite surface protein 4 (MSP4)-specific single-chain Fv protein

(scFv) inhibits P. falciparum growth in vitro [59]. Targeted GzmB-based drugs that can be delivered selectively into infected RBCs could potentially be used to treat malaria and other intercellular parasites. However, they would likely be too expensive to be practical in the countries where malaria is endemic. Combined therapy with drugs targeting multiple GzmB substrates may be used as an alternative.

Reference 1. Yeung, S., Malaria-Update on Antimalarial Resistance and Treatment Approaches. Pediatr Infect Dis J, 2018. 37(4): p. 367-369. 2. Beeson, J.G., et al., Plasmodium vivax malaria: challenges in diagnosis, treatment and elimination. Pediatr Infect Dis J, 2015. 34(5): p. 529-31. 3. Dondorp, A.M., et al., Artemisinin resistance in Plasmodium falciparum malaria. N Engl J Med, 2009. 361(5): p. 455-67. 4. Dogovski, C., et al., Targeting the cell stress response of Plasmodium falciparum to overcome artemisinin resistance. PLoS Biol, 2015. 13(4): p. e1002132. 5. Mbengue, A., et al., A molecular mechanism of artemisinin resistance in Plasmodium falciparum malaria. Nature, 2015. 520(7549): p. 683-7. 6. Kaslow, D.C. and S. Biernaux, RTS,S: Toward a first landmark on the Malaria Vaccine Technology Roadmap. Vaccine, 2015. 33(52): p. 7425-32. 7. Ishizuka, A.S., et al., Protection against malaria at 1 year and immune correlates following PfSPZ vaccination. Nat Med, 2016. 22(6): p. 614-23. 8. Kebaier, C., T. Voza, and J. Vanderberg, Kinetics of mosquito-injected Plasmodium sporozoites in mice: fewer sporozoites are injected into sporozoite-immunized mice. PLoS Pathog, 2009. 5(4): p. e1000399. 9. Cowman, A.F. and B.S. Crabb, Invasion of red blood cells by malaria parasites. Cell, 2006. 124(4): p. 755-66. 10. Casares, S. and T.L. Richie, Immune evasion by malaria parasites: a challenge for vaccine development. Curr Opin Immunol, 2009. 21(3): p. 321-30. 11. Izak, D., J. Klim, and S. Kaczanowski, Host-parasite interactions and ecology of the malaria parasite-a bioinformatics approach. Brief Funct Genomics, 2018. 17(6): p. 451- 457. 12. Soulard, V., et al., Plasmodium falciparum full life cycle and Plasmodium ovale liver stages in humanized mice. Nat Commun, 2015. 6: p. 7690. 13. Marsh, K. and S. Kinyanjui, Immune effector mechanisms in malaria. Parasite Immunol, 2006. 28(1-2): p. 51-60. 14. Boura, M., et al., The hemozoin conundrum: is malaria pigment immune-activating, inhibiting, or simply a bystander? Trends Parasitol, 2013. 29(10): p. 469-76. 15. Turner, L., et al., Severe malaria is associated with parasite binding to endothelial protein C receptor. Nature, 2013. 498(7455): p. 502-5. 16. Chan, J.A., F.J. Fowkes, and J.G. Beeson, Surface antigens of Plasmodium falciparum- infected erythrocytes as immune targets and malaria vaccine candidates. Cell Mol Life Sci, 2014. 71(19): p. 3633-57. 17. Cavanagh, D.R., et al., A longitudinal study of type-specific antibody responses to Plasmodium falciparum merozoite surface protein-1 in an area of unstable malaria in Sudan. J Immunol, 1998. 161(1): p. 347-59. 18. Walther, M., et al., Distinct roles for FOXP3 and FOXP3 CD4 T cells in regulating cellular immunity to uncomplicated and severe Plasmodium falciparum malaria. PLoS Pathog, 2009. 5(4): p. e1000364. 19. Illingworth, J., et al., Chronic exposure to Plasmodium falciparum is associated with phenotypic evidence of B and T cell exhaustion. J Immunol, 2013. 190(3): p. 1038-47. 20. Horne-Debets, J.M., et al., PD-1 dependent exhaustion of CD8+ T cells drives chronic malaria. Cell Rep, 2013. 5(5): p. 1204-13.

21. McGilvray, I.D., et al., Nonopsonic monocyte/macrophage phagocytosis of Plasmodium falciparum-parasitized erythrocytes: a role for CD36 in malarial clearance. Blood, 2000. 96(9): p. 3231-40. 22. Schwarzer, E. and P. Arese, Phagocytosis of malarial pigment hemozoin inhibits NADPH-oxidase activity in human monocyte-derived macrophages. Biochim Biophys Acta, 1996. 1316(3): p. 169-75. 23. Junqueira, C., et al., Cytotoxic CD8(+) T cells recognize and kill Plasmodium vivax- infected reticulocytes. Nat Med, 2018. 24(9): p. 1330-1336. 24. Martinvalet, D., P. Zhu, and J. Lieberman, Granzyme A induces caspase-independent mitochondrial damage, a required first step for apoptosis. Immunity, 2005. 22(3): p. 355- 70. 25. Sutton, V.R., et al., Caspase activation by granzyme B is indirect, and caspase autoprocessing requires the release of proapoptotic mitochondrial factors. Immunity, 2003. 18(3): p. 319-29. 26. Adrain, C., B.M. Murphy, and S.J. Martin, Molecular ordering of the caspase activation cascade initiated by the cytotoxic T lymphocyte/natural killer (CTL/NK) protease granzyme B. J Biol Chem, 2005. 280(6): p. 4663-73. 27. Dotiwala, F., et al., Granzyme B Disrupts Central Metabolism and Protein Synthesis in Bacteria to Promote an Immune Cell Death Program. Cell, 2017. 171(5): p. 1125-1137 e11. 28. Dotiwala, F., et al., Killer lymphocytes use granulysin, perforin and granzymes to kill intracellular parasites. Nat Med, 2016. 22(2): p. 210-6. 29. Bots, M. and J.P. Medema, Granzymes at a glance. J Cell Sci, 2006. 119(Pt 24): p. 5011- 4. 30. Parham, P., MHC class I molecules and KIRs in human history, health and survival. Nat Rev Immunol, 2005. 5(3): p. 201-14. 31. Kruse, P.H., et al., Natural cytotoxicity receptors and their ligands. Immunol Cell Biol, 2014. 92(3): p. 221-9. 32. Newman, K.C., et al., Cross-talk with myeloid accessory cells regulates human natural killer cell interferon-gamma responses to malaria. PLoS Pathog, 2006. 2(12): p. e118. 33. Artavanis-Tsakonas, K. and E.M. Riley, Innate immune response to malaria: rapid induction of IFN-gamma from human NK cells by live Plasmodium falciparum-infected erythrocytes. J Immunol, 2002. 169(6): p. 2956-63. 34. Korbel, D.S., et al., Killer Ig-like receptor (KIR) genotype predicts the capacity of human KIR-positive CD56dim NK cells to respond to pathogen-associated signals. J Immunol, 2009. 182(10): p. 6426-34. 35. Artavanis-Tsakonas, K., et al., Activation of a subset of human NK cells upon contact with Plasmodium falciparum-infected erythrocytes. J Immunol, 2003. 171(10): p. 5396- 405. 36. Baratin, M., et al., Natural killer cell and macrophage cooperation in MyD88-dependent innate responses to Plasmodium falciparum. Proc Natl Acad Sci U S A, 2005. 102(41): p. 14747-52. 37. Arora, G., et al., NK cells inhibit Plasmodium falciparum growth in red blood cells via antibody-dependent cellular cytotoxicity. Elife, 2018. 7. 38. Sherwood, A.M., et al., Deep sequencing of the human TCRgamma and TCRbeta repertoires suggests that TCRbeta rearranges after alphabeta and gammadelta T cell commitment. Sci Transl Med, 2011. 3(90): p. 90ra61.

39. Abeler-Dorner, L., et al., Butyrophilins: an emerging family of immune regulators. Trends Immunol, 2012. 33(1): p. 34-41. 40. Sandstrom, A., et al., The intracellular B30.2 domain of butyrophilin 3A1 binds phosphoantigens to mediate activation of human Vgamma9Vdelta2 T cells. Immunity, 2014. 40(4): p. 490-500. 41. Ho, M., et al., Increased gamma delta T cells in acute Plasmodium falciparum malaria. Immunol Lett, 1990. 25(1-3): p. 139-41. 42. Troye-Blomberg, M., et al., Human gamma delta T cells that inhibit the in vitro growth of the asexual blood stages of the Plasmodium falciparum parasite express cytolytic and proinflammatory molecules. Scand J Immunol, 1999. 50(6): p. 642-50. 43. Dantzler, K.W. and P. Jagannathan, gammadelta T Cells in Antimalarial Immunity: New Insights Into Their Diverse Functions in Protection and Tolerance. Front Immunol, 2018. 9: p. 2445. 44. Liu, C., et al., Vgamma9Vdelta2 T cells proliferate in response to phosphoantigens released from erythrocytes infected with asexual and gametocyte stage Plasmodium falciparum. Cell Immunol, 2018. 334: p. 11-19. 45. Guenot, M., et al., Phosphoantigen Burst upon Plasmodium falciparum Schizont Rupture Can Distantly Activate Vgamma9Vdelta2 T Cells. Infect Immun, 2015. 83(10): p. 3816- 24. 46. Wrobel, P., et al., Lysis of a broad range of epithelial tumour cells by human gamma delta T cells: involvement of NKG2D ligands and T-cell receptor- versus NKG2D- dependent recognition. Scand J Immunol, 2007. 66(2-3): p. 320-8. 47. Gertner, J., et al., Human gammadelta T lymphocytes strip and kill tumor cells simultaneously. Immunol Lett, 2007. 110(1): p. 42-53. 48. Costa, G., et al., Control of Plasmodium falciparum erythrocytic cycle: gammadelta T cells target the red blood cell-invasive merozoites. Blood, 2011. 118(26): p. 6952-62. 49. Farouk, S.E., et al., Gamma delta T cells inhibit in vitro growth of the asexual blood stages of Plasmodium falciparum by a granule exocytosis-dependent cytotoxic pathway that requires granulysin. Eur J Immunol, 2004. 34(8): p. 2248-56. 50. Morita, C.T., et al., Nonpeptide antigens, presentation mechanisms, and immunological memory of human Vgamma2Vdelta2 T cells: discriminating friend from foe through the recognition of prenyl pyrophosphate antigens. Immunol Rev, 2007. 215: p. 59-76. 51. Booden, T. and R.W. Hull, Nucleic acid precursor synthesis by Plasmodium lophurae parasitizing chicken erythrocytes. Exp Parasitol, 1973. 34(2): p. 220-8. 52. Madrid, D.C., et al., Plasmodium falciparum purine nucleoside phosphorylase is critical for viability of malaria parasites. J Biol Chem, 2008. 283(51): p. 35899-907. 53. Walch, M., et al., Cytotoxic Cells Kill Intracellular Bacteria through Granulysin- Mediated Delivery of Granzymes. Cell, 2015. 161(5): p. 1229. 54. Estebanez-Perpina, E., et al., Crystal structure of the caspase activator human granzyme B, a proteinase highly specific for an Asp-P1 residue. Biol Chem, 2000. 381(12): p. 1203-14. 55. Dziekan, J.M., et al., Identifying purine nucleoside phosphorylase as the target of quinine using cellular thermal shift assay. Sci Transl Med, 2019. 11(473). 56. Kicska, G.A., et al., Purine-less death in Plasmodium falciparum induced by immucillin- H, a transition state analogue of purine nucleoside phosphorylase. J Biol Chem, 2002. 277(5): p. 3226-31.

57. Mather, M.W. and A.B. Vaidya, Mitochondria in malaria and related parasites: ancient, diverse and streamlined. J Bioenerg Biomembr, 2008. 40(5): p. 425-33. 58. Sakata-Kato, T. and D.F. Wirth, A Novel Methodology for Bioenergetic Analysis of Plasmodium falciparum Reveals a Glucose-Regulated Metabolic Shift and Enables Mode of Action Analyses of Mitochondrial Inhibitors. ACS Infect Dis, 2016. 2(12): p. 903-916. 59. Kapelski, S., et al., Antimalarial activity of granzyme B and its targeted delivery by a granzyme B-single-chain Fv fusion protein. Antimicrob Agents Chemother, 2015. 59(1): p. 669-72.