THE ROLE OF RHOMBOID PROTEASES AND A OOCYST CAPSULE PROTEIN IN MALARIA PATHOGENESIS AND PARASITE DEVELOPMENT PRAKASH SRINIVASAN

Transcription

1 THE ROLE OF RHOMBOID PROTEASES AND A OOCYST CAPSULE PROTEIN IN MALARIA PATHOGENESIS AND PARASITE DEVELOPMENT BY PRAKASH SRINIVASAN Submitted in partial fulfillment of the requirements For the degree of Doctor of Philosophy Thesis Advisor: Prof. Marcelo Jacobs-Lorena Department of Genetics CASE WESTERN RESERVE UNIVERSITY August, 2007

2 CASE WESTERN RESERVE UNIVERSITY SCHOOL OF GRADUATE STUDIES We hereby approve the dissertation of candidate for the Ph.D. degree *. (signed) (chair of the committee) (date) *We also certify that written approval has been obtained for any proprietary material contained therein.

3 TABLE OF CONTENTS Table of Contents 1 List of Tables 2 List of Figures 4 Acknowledgements 5 Abstract 7 CHAPTER 1: Introduction and Research Objectives 9 Introduction 10 Malaria: History and Facts 10 Discovery of Mosquitoes as vectors 10 Malaria: Life Cycle 12 Life cycle in the vertebrate host 12 Life cycle in the mosquito 15 Sporozoite invasion of the liver 29 Study of gene function in parasites 32 Research Objectives 34 CHAPTER 2: Analysis of Plasmodium and Anopheles Transcriptomes during Oocyst Differentiation 37 CHAPTER 3: PbCap380, a novel Plasmodium Oocyst Capsule Protein is Essential for Parasite Survival in the Mosquito 61 CHAPTER 4: Distinct Roles for Rhomboid1 and Rhomboid3 Proteins in Malaria Pathogenesis and Plasmodium Development 82 CHAPTER 5: Conclusions and Future Directions 107 1

4 Conclusion 108 Future Directions 111 Bibliography 118 2

5 LIST OF TABLES CHAPTER 2 Table CHAPTER 3 Table CHAPTER 4 Table

6 LIST OF FIGURES CHAPTER 1 Figure Figure Figure CHAPTER 2 Figure Figure Figure Figure Figure Figure CHAPTER 3 Figure Figure Figure Figure Figure

7 CHAPTER 4 Figure Figure Figure Figure Figure Figure Figure Figure Figure

8 Acknowledgements I have had the pleasure and privilege of working with a number of people during the course of my study. I would like to first thank my mentor, Dr. Marcelo Jacobs- Lorena for his support and guidance through all these years. His passion for science and an eagerness to face new challenges have always been motivating. His willingness to let me pursue my work independently has helped me develop my scientific knowledge and has given me the confidence to follow my scientific dreams. My sincere thanks to my committee members, Dr. Bruce Lamb, Dr. Peter Harte, Dr. Pete Zimmerman and Dr. Edward Stavnezer for their constructive criticisms and invaluable suggestions. I also thank the members of the MJL lab, past and present, for their support and for making the lab meetings a fun time (for the most part!). I would like to thank Anil Ghosh and Abraham Eappen for their help, particularly during my initial days in the lab. I take this opportunity to thank all my teachers for their encouragement and support, especially Dr. Patrick Gomez and Dr. Yogesh Shouche. I thank my friends Sumi, Sujit, Anirudh, Cristina and Jessica for their support and the fun memories. Special thanks to my dear friend Minnie for always being there, for the 6

9 stimulating discussions, the delicious meals (and the one of a kind tea!) and wonderful memories. I would not be here if not for the unfathomable love and incredible support from my parents Srinivasan and Vimala. They have always believed in me and encouraged me to follow my dreams. The love and affection of my wonderful sister has been a source of energy. I thank them from the bottom of my heart. I m forever indebted to my late grand parents, whose love, support and confidence in me have made me and still make me a better person in life. Finally, thanks to my dearest friend and my best half, Susham, with whom I have shared the best and worst times of my graduate life. I can definitely say without a doubt that she made graduate school a lot easier with her patience, support and love and has helped broaden my scientific thinking and believe in my abilities. 7

10 The Role of Rhomboid Proteases and a Oocyst Capsule protein in Malaria Pathogenesis and Parasite Development Abstract by PRAKASH SRINIVASAN Plasmodium, the etiological agent of malaria causes more than a million deaths mostly in children under the age of five, and nearly 500 million people suffer from this disease every year. The lack of an effective vaccine and the spread of resistance against currently used drugs underscores the need to identify new targets and approaches to control this disease. Plasmodium life cycle takes place in two hosts, namely, the vertebrate and the mosquito. The sexual stages of the parasite occur in the mosquito. This absolute dependence on the mosquito for parasite represents a potential weak link that could be exploited. A better understanding of parasite development in the mosquito may identify new targets for intervention. In this work, I have attempted to address some of the questions pertaining to specific developmental events that take place during parasite development in the midgut. Soon after fertilization the motile ookinete that is formed inside the blood meal transforms 8

11 into a sessile oocyst. The parasite then enlarges in size and a single oocyst can produce thousands of sporozoites, the infective forms. The factors that regulate oocyst differentiation are not well understood. This study identifies several novel genes that are expressed during oocyst development. During parasite development in the mosquito, a drastic reduction in number occurs as the parasite transforms from one stage to the other. However, once oocysts are formed they appear to be resistant against the mosquito defenses. In this study we identify an important function for a oocyst capsule protein in parasite survival in the mosquito. Finally, this work also sheds light on the role of two rhomboid family serine proteases in parasite development and malaria pathogenesis. Overall, this study advances our understanding of Plasmodium development in the mosquito as well as in the vertebrate host and identifies potential targets for interfering with parasite transmission and malaria pathogenesis. 9

12 CHAPTER 1 Introduction and Research Objectives 10

13 Introduction Malaria: History and Facts Malaria is one of the oldest and the most debilitating diseases known to man. Nearly 50% of the world s population is at risk and an estimated 1 to 2.7 million people die of malaria every year, mostly children under the age of five (Breman et al. 2001, Greenwood and Mutabingwa 2002). Malaria is thought to have been prevalent as early as 2700 B.C. Two of the most widely used drugs of recent times, artemisinin and quinine were in fact derived from plants used in ancient Chinese and South American medicine to treat malaria-like symptoms. However the causative agent, Plasmodium, a protozoan parasite was discovered only a century ago. Charles Lavern in 1880 was the first to identify parasites in the blood of a patient with symptoms of malaria. Since then four species of Plasmodium: P. falciparum, P. vivax, P. ovale and P. malariae that infect humans have been identified. Malaria: Discovery of mosquitoes as vectors A major breakthrough in the control of malaria came when Ronald Ross (1897) discovered mosquitoes as vectors for malaria transmission. Ross was very well aware of the importance of this discovery and this is reflected in his poem, This day designing God Hath put into my hand 11

14 A wondrous thing; and God Be praised. At his command I have found thy cunning seeds Oh million-murdering Death, I know that this little thing A million men will save Oh death where is thy sting? Thy victory oh grave? Subsequently, female Anopheles mosquitoes were identified as the sole vectors of human malaria transmission. This absolute dependence on the mosquitoes for parasite transmission reveals a potential weak link that has been exploited by the use of insecticides (mainly DTT and pyrithroids) to reduce malaria. The use of drugs against the parasite and insecticides against mosquito resulted in a drastic reduction of malaria. However, emergence of insecticide resistant mosquitoes (Hemingway et al. 2004) and multiple drug resistant parasites (Sidhu et al. 2002, Price et al. 2004, Valderramos and Fidock 2006) has led to a global crisis in controlling this disease. More information on the milestones and history of malaria are recounted by Harrison (1978) and Desowitz (1991). The following web sites also provide useful information on malaria

15 Malaria: Life cycle The life cycle of the malaria parasite involves development in two different hosts: vertebrate host (exoerythrocytic and erythrocytic forms (also called asexual or blood stage)) and mosquito vector (sexual stage) (Fig.1.1 and Fig.1.2). Life cycle in the vertebrate host Infection begins soon after the transfer of sporozoites (infective forms) into a naïve host by an infected mosquito (Fig.1.1). The sporozoite carried through the blood circulation infects the liver (exoerythrocytic stage). After 4-6 days, a single infected hepatocyte can produce nearly 30,000 merozoites Upon rupture of the infected hepatocyte, merozoites released into the blood stream invade the red blood cells (RBCs) and start the erythrocytic stage of development. P. falciparum is the most widespread and the deadliest of the four human malaria species. A unique feature of P. vivax, P. ovale and P.cynomolgi (monkey malaria) is the relapse of malaria. This is a process whereby exoerythrocytic forms of the parasites stay dormant (hypnozoites) for months or sometimes, even years without any symptom (Cogswell 1992) and are reactivated by yet unknown mechanisms. Within the RBC, the merozoite undergoes a series of developmental changes: ring stage, trophozoite stage, and schizonts. During schizogony, the parasite undergoes DNA replication, resulting in the formation of 8-24 merozoites. Merozoites released 13

16 Fig.1.1. Life cycle of Plasmodium in the vertebrate host. The malaria parasite life cycle involves two hosts. During a blood meal, a malaria-infected female Anopheles mosquito inoculates sporozoites into the human host (1). Sporozoites infect liver cells (2) and mature into schizonts (3), which rupture and release merozoites (4). (Of note, in P. vivax and P. ovale a dormant stage [hypnozoites] can persist in the liver and cause relapses by invading the bloodstream weeks, or even years later.) After this initial replication in the liver (A-exoerythrocytic schizogony), the parasites undergo asexual multiplication in the erythrocytes (Berythrocytic schizogony). Merozoites infect red blood cells (5). The ring stage trophozoites mature into schizonts, which rupture releasing merozoites (6). Some parasites differentiate into sexual erythrocytic stages (gametocytes, 7). Blood stage parasites are responsible for the clinical manifestations of the disease. The gametocytes, male (microgametocytes) and female (macrogametocytes), are ingested by a female Anopheles mosquito during a blood meal (8). Adapted from the CDC web site, 14

17 upon schizont rupture invade new RBCs and continue this asexual cycle. This cycle of merozoite invasion of RBCs and subsequent release of merozoites from the infected RBC occurs every 48 h in P. falciparum, P. vivax and P. ovale while it takes 72 h in P. malariae. The clinical manifestation of the disease, chills and fever, are associated with the rupture of schizonts and the release of new merozoites. As this cycle continues, more and more RBCs become infected and eventually lead to malarial anemia. The surface of the infected RBC is modified by the addition of parasite-derived proteins (Smith and Craig 2006). In severe cases, this leads to cerebral malaria caused possibly by a breach in the blood-brain barrier and sequestration of parasitized red blood cells and host leukocytes within the cerebral microvasculature (Medana and Turne 2006). Commitment to sexual differentiation (gametocytogenesis). As the parasites go through the asexual stage development, a small fraction commit to sexual development. There has been much debate about the actual point of commitment. Bruce and colleagues have shown that merozoites emerging from a single schizont invade RBCs and develop into either ring stages or into gametocytes (Bruce et al. 1990). Subsequently it was shown that all the merozoites from one schizont commit to forming either male or female gametocyte (Smith et al. 2000, Silvestrini et al. 2000). This suggests that commitment to gametocytogenesis occurs in the preceding asexual generation. 15

18 Gametocyte development. In most Plasmodium species, gametocytogenesis is of longer duration than asexual development. For instance in P. falciparum, it is 9-12 days compared to h for asexual stage development. Morphologically, the immature gametocyte resembles the trophozoite. At this stage the male and female gametocytes cannot be distinguished. In the case of P. falciparum, they can be distinguished 1-3 days after development by their elongated plasma membrane. Gametocyte maturation can be divided morphologically into five stages (stage I stage V) during which an inner pellicular membrane vacuole and microtubules form beneath the plasma membrane (Sinden 1983). This gives the gametocyte its distinct shape and symmetry. However, in rodent parasites such as P. berghei, the gametocyte is round and can be distinguished from the trophozoites by the presence of nuclear granules. Life cycle in the mosquito Mosquitoes are an absolute requirement for malaria transmission and represent a potential weak link in the life cycle of the parasite. Development of Plasmodium in the mosquito is a highly complex process that involves transformation into six different forms (Fig.1.2). When a female Anopheles mosquito bites an infected host, the gametocytes are taken up along with the blood meal. Within minutes, gametocytes in the mosquito midgut differentiate into male and female gametes. Each male gametocyte (microgametocyte) gives rise to eight male gametes. Within hours, fertilization of male and female gamete forms the zygote, which then transforms into 16

19 an elongated, motile ookinete. The ookinete migrates actively in the blood meal and traverses through the midgut epithelium. In doing so, it has to first penetrate a chitinrich peritrophic matrix (PM). The PM surrounds the blood meal and is thought to protect the mosquito from microorganisms entering the midgut (Abraham and Jacobs- Lorena 2004). In order for the parasite to reach the midgut epithelium, it secretes chitinase, an enzyme that cleaves the chitin present in the PM (Huber et al. 1991). The ookinete then invades the midgut epithelium. During this process the invaded cells undergo apoptosis and are detached from the epithelium into the midgut lumen (Han et al. 2000). However, ookinetes that exit these cells and reach the basal lamina before being excluded from the midgut. Reaching the basal lamina leads to an arrest in kinesis and triggers the transformation into a spherical oocyst. The oocyst remains sandwiched between the epithelium and the basal lamina for days (the longest duration of any parasite stage). During oocyst development, the parasite grows in size and undergoes numerous mitotic divisions resulting in the formation of thousands of sporozoites. Upon maturation, these sporozoites are released into the hemocoel. Of all the tissues they come in contact with, the sporozoites selectively invade the salivary glands. Transmission occurs when the infected mosquito bites a vertebrate host in the process transferring the infective sporozoites. This absolute requirement of the mosquito for transmission makes the sexual stages of the parasite attractive targets for interruption 17

20 strategies. For instance antigens expressed on the surface of gametes and ookinetes have been used as transmission blocking vaccines in in vitro assays and animal models (Quakyi et al. 1987, Barr et al. 1991, Kaslow 2002). Fig.1.2 Plasmodium life cycle in the mosquito. Sexual development begins in the mosquito midgut lumen with differentiation of gametocytes into male and female gametes, followed by fertilization and differentiation of the resulting zygotes into motile ookinetes. Ookinetes migrate to, and invade the midgut epithelium and then lodge underneath the basal lamina. Contact with the basal laminai appears to trigger differentiation of ookinetes into oocysts. The thousands of sporozoites that are formed inside the mature oocyst are released into the hemocoel. About 20% of the released sporozoites invade the salivary gland and are transmitted when the mosquito feeds on another individual. The approximate time line for each differentiation step is indicated (from Ghosh et al. 2003). Specific proteins are expressed at different stages of the parasite development. An overview of the various processes involved and some of the underlying mechanisms are described below. 18

21 Ookinete formation. Within minutes after entering the mosquito midgut, the gametocytes escape from the erythrocytes and form male and female gametes. Each male gametocyte gives rise to eight male gametes. Fertilization takes places within the first couple of hours resulting in a diploid zygote. Two proteins (P48/45 and P230) have been shown to be essential for this process (Kocken et al. 1993, Williamson et al. 1993, van Dijk et al. 2001). The P48/45 protein is specific to male gametes and parasites lacking this protein cannot fertilize. However, when female gametes from mutant parasites are mixed with wild type male gametes, they form viable ookinetes and infect the mosquito (van Dijk et al. 2001). Transformation of the spherical zygote into a banana-shaped motile ookinete varies between Plasmodium species and is temperature dependent. Maturation of P. berghei ookinetes in Anopheles stephensi occurs in h at 19 o C, while P. yoelii and P. falciparum develop at 26 o C and take h and h respectively. Ookinete invasion of the midgut epithelium. A number of proteins are expressed specifically in the ookinetes. Some of them such as P25 and P28 are transcribed during gametocytogenesis in the vertebrate host but are translated after gamete differentiation in the mosquito. Some others such as chitinase, circumsporozoite and TRAP related protein (CTRP), von Willebrand factor A domain-related protein (WARP) and secreted ookinete adhesive protein (SOAP) are transcribed following gametocyte activation (Huber et al. 1991, Dessens et al. 1999, Yuda et al. 2001). 19

22 The early stages of sexual development in the mosquito take place within the blood meal that is surrounded by a chitin-rich peritrophic matrix (PM). The ookinete is sensitive to mosquito proteases (Gass and Yeates 1979). In order to access the epithelium, ookinetes have to travel past the PM. To do so, the parasite secretes chitinase, a enzyme that cleaves chitin proteins of the PM (Shahabuddin et al. 1993). Studies have shown that a mosquito trypsin secreted into the midgut lumen activates Pgchitinase (P. gallinaceum, avian parasite) (Shahabuddin et al. 1993). Trypsin inhibitors when added to the blood meal blocked ookinete invasion (Shahabuddin et al. 1993). Parasites lacking chitinase are reduced in their ability to form oocysts, presumably because of inability of the ookinetes to effeciently move past the PM and thereby getting eliminated from the midgut (Tsai et al. 2001, Dessens et al. 2001). After crossing the PM, the ookinete actively invades the midgut epithelium employing gliding motility (see below). CTRP a type I transmembrane protein containing multiple adhesive motifs (Yuda et al. 1999) is stored in micronemes and secreted from the apical end of the ookinete as it migrates (Yuda et al. 1999). CTRP is a member of the thrombospondin-domain protein family. In sporozoites a related member, TRAP (thrombospondin related adhesive protein) is involved in gliding motility (Sultan et al. 1997). Similarly, CTRP knockout parasites are non-motile and are unable to invade the midgut epithelium implying CTRP is required for motility (Dessens et al. 1999, Templeton et al. 2000). In addition, the presence of adhesive domains in the extracellular region of CTRP indicates that it could also function in attachment to certain mosquito receptors. 20

23 A major rearrangement in the surface proteins occurs after fertilization. All known pre-existing surface proteins are shed and new proteins are made. Two of the dominant proteins are P25 and P28 that are GPI anchored to the membrane (Tomas et al. 2001). Both proteins contain multiple epidermal growth factor-like (EGF) domains and interact with the EGF domains present in laminin gamma1. Phenotypic analysis of P25 andp28 single and double knockout ookinetes suggest these proteins have mutually redundant functions including protection against mosquito midgut proteases (Tomas et al. 2001, Saxena et al. 2006). The single knockout parasites formed oocyst in numbers similar to wild type parasites, while the number of oocysts in double knockout parasites was reduced by as much as 99.5% (Tomas et al. 2001). WARP, a von Willebrand domain-a containing protein is thought to be present on the surface of the ookinete and early oocyst (Yuda et al. 2001, Abraham et al. 2004). Antibodies against WARP when fed to mosquitoes inhibit ookinete to oocyst transformation (Abraham et al. 2004). The presence of the adhesive domain indicates a role in binding to a mosquito receptor. The gene for SOAP encodes a predicted secreted protein with a modular structure composed of two unique cysteine-rich domains (Dessens et al. 2003). In yeast-twohybrid assays, SOAP has been shown to interact with mosquito laminin. Laminin is a very abundant protein of the basal lamina. SOAP knockout parasites formed significantly fewer oocysts indicating that interaction of SOAP with laminin could be one of the signals that trigger ookinete to oocyst transformation (Dessens et al. 2003). 21

24 Plasmodium berghei scavenger receptor-like protein (SR) is composed of a unique combination of metazoan protein domains, namely: (i) scavenger receptor cysteine rich domain (SRCR); (ii) pentraxin domain (PTX); (iii) polycystine-1, lipoxygenase, alpha toxin domain (LH2/PLAT); (iv) Limulus clotting factor C, Coch-5b2 and Lgl1 domain (LCCL). These domains have been previously associated with immune recognition/activation and lipid/protein adhesion interactions at the cell surface. PbSR protein is expressed in sporozoites and mutants lacking this protein form normal numbers of oocysts but the oocysts are devoid of sporozoites (Claudianos et al. 2002). More recently Pradel et al (2004) described a family of 5 proteins in P. falciparum, which they called PfCCp1-5. PfCCp3 is the orthologue of PbSR. PfCCp1-3 are expressed in gametocytes and are secreted during gametogenesis. However, PfCCp2 and PfCCp3 knockout parasites in contrast to PbSR formed fully developed sporozoite-containing oocysts but were defective in transiting to the salivary glands (Pradel et al. 2004). This difference between P. berghei and P. falciparum could represent potential differences between the different Plasmodium species. More recently a family of two micronemal proteins containing membrane-attack complex and perforin (MACPF)-related domain was identified in Plasmodium (Kadota et al. 2004, Ishino et al. 2005). These domains are present in pore forming proteins. PbMAOP (membrane attack ookinete protein) is expressed specifically in the ookinete (Kadota et al. 2004) while PbSPECT (sporozoite microneme protein essential for cell traversal) is expressed in sporozoites (see below). Mutant PbMAOP ookinetes were unable to invade the midgut epithelium and seemed to accumulate at 22

25 the surface (Kadota et al. 2004). A similar role for PbSPECT is seen during sporozoite invasion of the hepatocyte (Ishino et al. 2005). This suggests that a conserved mechanism for membrane rupture involving MACPF-related proteins is used in different host invasive stages of the malarial parasite. The extracellular nature and its restricted expression in the mosquito midgut stages of Plasmodium make some of the above-described proteins a potential transmissionblocking antibody target and vaccine candidates. Indeed, P25 is in clinical trials and others including WARP and SOAP have shown promise in animal models. These results indicate that sexual stage antigens could prove useful as a component in a multiantigen-based malaria transmission blocking vaccine, and should stimulate further research on this topic. Ookinete-to-Oocyst transformation. Upon traversing the midgut epithelium, the ookinete comes in contact with the basal lamina. Contact with the basal lamina leads to an arrest in motility and triggers the rounding up of the elongated ookinete resulting in a non-motile oocyst. The molecular basis for this transformation is not well understood. Ookinete surface proteins P25, P28, SOAP and CTRP have been shown to interact with certain basal lamina components such as gamma 1 laminin (Vlachou et al. 2001, Dessens et al. 2003, Mahairaki et al. 2005). However it is not clear if these interactions occur in vivo and even if they do, whether this interaction is sufficient to signal the transformation. There could be other components in addition to laminin that interact with the ookinetes. Culture methods developed (Al-Olayan et 23

26 al. 2002) to grow oocysts in vitro indicated extracellular matrix components to be necessary for successful transformation of ookinete to oocyst. Subsequently, other studies have shown that oocyst development can take place in vitro in the presence of cells and in the absence of basal lamina components (Siden-Kiamos et al. 2000). Therefore the requirement of parasite surface proteins to interact with mosquito basal lamina components as a trigger for oocyst formation has not been fully established. Oocyst development and sporozoite formation. At the molecular level, the oocyst is the least understood form of the parasite. The newly formed oocyst (4 µm) remains sandwiched between the midgut epithelium and the basal lamina. Over the next days the oocyst rapidly increases in size (40-60 µm). One of the important functions of the oocyst is to increase the parasite numbers that have gone through a bottleneck in the process of reaching the basal lamina. After fertilization, the resulting zygote and ookinete contain 4 haploid genomes as a result of meiosis. The oocyst as such inherits the 4 haploid copies and later undergoes 11 synchronized mitotic divisions yielding haploid nuclei. Towards the end of oocyst development, the plasma membrane retracts from the capsule and invaginates into the oocyst cytoplasm (Thathy et al. 2002). This process subdivides the oocyst into several cytoplasmic islands called the sporoblasts. Subsequently, double membranes are formed below the plasma membrane at specific sites of sporozoite budding. Various organelles such as ER, Golgi and other sporozoite apical organelles are distributed throughout the cytoplasm (Fig.1.3A). Sporozoite formation begins days after the infective blood meal. Budding occurs at specific electron dense sites along the periphery of the 24

27 sporoblast membrane (Thathy et al. 2002). The budding parasite inherits a single nucleus along with other organelles to form the sporozoite. The sporozoite-filled oocyst then ruptures into the mosquito hemocoel releasing the infective forms (Sinden 1974). Soon after the formation of oocyst, the parasite secretes an amorphous capsule that surrounds the entire oocyst (Fig.1.3B). The capsule separates the parasite from the mosquito tissues. Electron microscopy studies show that the capsule is in direct contact with the mosquito tissues (Fig.1.3A). A parasite-specific transglutaminase activity was reported (Adini et al. 2001). It is suggested that proteins in the capsule could be cross-linked with one another as well as to some mosquito-derived proteins. However, the precise nature/composition of the capsule is not known. As discussed earlier, the parasite undergoes a drastic reduction in numbers as it transits from the blood meal towards the basal lamina. This is mainly due to the efficient recognition of the parasite by the mosquito immune system resulting in the killing of parasites. The effect of the mosquito defenses against the developing oocyst remains unclear. It is possible that the non-motile oocyst is not detected by the mosquito, whereas the parasites invading the midgut lumen are easily recognized. Another possibility is that the oocyst capsule protects the parasite from the mosquito defense molecules. The capsule could keep the host immune molecules from reaching the parasite or through its interactions with the basal lamina be coated with mosquito molecules thereby remaining invisible. Therefore if one can interfere with the formation of the oocyst 25

28 capsule it may be possible to eliminate the last few parasites before the massive production of infective sporozoites can occur. A B Fig.1.3. Electron microscopy of developing oocyst. A, Developing oocyst that has undergone karyokinesis resulting in a multi-nucleated (N) parasite. The oocyst cytoplasm also contains organelles such as the endoplasmic reticulum (ER). B, Mature oocyst surrounded by the capsule (C) in which sporoblasts have formed (Sb) and are undergoing sporozoite budding (S). The molecular mechanisms that regulate oocyst development and sporozoite formation are not well understood. Circumsporozoite protein (CSP), a thrombospondin type I domain containing GPI anchored surface protein is expressed before sporozoite formation. Gene knockout studies reveal an arrest in sporozoite budding (Menard et al. 1997). In wild type parasites, the emerging inner membranes where budding is initiated are restricted to small areas. In CSP(-) parasites, the budding sites become spread over larger areas of the plasma membrane thereby preventing sporozoite budding (Thathy et al. 2002). 26

29 Recent studies have shown that a transmembrane domain cannot substitute the GPIanchor, suggesting that the GPI anchor is required for sporogenesis (Wang et al. 2005a). Mutating a conserved motif of positively charged amino acids in the CS protein resulted in the oocysts forming sporozoites (Wang et al. 2005b). However, these sporzoites could not exit the oocyst. Despite these interesting observations, the molecular basis of how CSP regulates these processes is not understood. PxSR/PxCCP3 is a predicted secreted protein described above. The P. berghei orthologue PbSR is expressed in sporozoites (Delrieu et al. 2002, Claudianos et al. 2002) in contrast to PfCCP3, which is expressed in gametocytes (Pradel et al. 2004). PbSR(-) parasites develop into oocysts but do not form sporozoites. This is in contrast to PfCCP3(-) parasites, which does form sporozoites but are not released from the oocyst. The reason for the differences between the two parasite species is not understood. At the end of sporozoite development, the oocyst is completely filled with sporozoites. The mechanism of sporozoite release from the oocyst is not well understood. A cysteine protease that is expressed specifically in oocyst sporozoites but not in salivary gland sporozoites was shown to be essential for sporozoite release (Aly and Matuschewski 2005). This suggests that sporozoite release is an active process involving at least one enzyme and not just by mechanical force resulting from the fully distended oocyst. 27

30 Sporozoite invasion of the salivary gland. Sporozoites that are released from the oocysts specifically invade the distal lateral and medial lobes of the salivary gland. Sporozoite invasion like the ookinete invasion is mediated by interactions between sporozoite surface ligand(s) and mosquito receptor(s). A few parasite ligands required for invasion have been described (see below), but their salivary gland receptors are not known. Antibodies raised against whole salivary glands identified Saglin, a surface protein as one of the receptors for sporozoite invasion (Brennan et al. 2000), but the parasite ligand is not known. Circumsporozoite protein (CSP), the major surface protein of sporozoites binds preferentially to the salivary glands and not to other mosquito tissues (Sidjanski et al. 1997). In addition, both the recombinant protein and a peptide from the N-terminal portion of CSP inhibits sporozoite invasion of the salivary glands (Myung et al. 2004). CSP contains two motifs (region I and region II), which are conserved among all Plasmodium species. Disruption of region II abolished the ability of sporozoites to invade salivary glands as well as infect the vertebrate host (Tewari et al. 2002). CSP also appears to provide the parasite with vector specificity. Rodent parasites that infect Anopheles mosquitoes when replaced with the CSP gene from avian parasite that infect Aedes mosquitoes, are able to form sporozoites but are unable to invade salivary glands (Tewari et al. 2005). 28

31 Thrombospondin related adhesive protein (TRAP) is type I transmembrane domain protein that contains von Willebrand A and thrombospondin type I domains (Robson et al. 1995). The short cytoplasmic tail is conserved in all Plasmodium species and interacts with the actin cytoskeleton through aldolase and myosin tail interacting protein (Buscaglia et al 2003, Bergman et al. 2003). Sporozoites with an altered TRAP cytoplasmic tail abolishe its contact with the actin cytoskeleton and render the parasite non-motile (Bhanot et al. 2003). The adhesive domains in TRAP are thought to bind to substrates and propel the parasite forward (Wengelnik et al. 1999). As the sporozoite moves forward TRAP is cleaved off and can be seen leaving a trail of TRAP and CSP (Stewart and Vanderberg 1991). The enzyme that cleaves TRAP is not known. Parasites lacking TRAP are unable to glide and do not invade the salivary glands and cannot infect the vertebrate host (Sultan et al. 1997). Only a few oocyst sporozoties have a small amount of TRAP on the surface while most of the TRAP protein is stored in micronemes (Bhanot et al. 2003). On the other hand almost all the salivary gland sporozoites have TRAP on their surface in addition to micronemal TRAP (Bhanot et al. 2003). MAEBL, a parasite surface protein initially thought to be asexual-stage protein is expressed in sporozoites (Srinivasan et al. 2004). Moreover, the protein appears to be localized inside the sporozoite and re-localized to the surface before invading the salivary glands (Srinivasan et al. 2004). MAEBL knockout parasites are unable to invade the salivary glands (Kariu et al. 2002) 29

32 Sporozoite invasion of the liver. When an infected mosquito feeds on a vertebrate host, the sporozoites are introduced in the dermis of the host. From the dermis, the sporozoites somehow find their way to a nearby blood vessel and reach the liver. The molecular cues that regulate this process are not known. Sporozoites from salivary glands are highly infective compared to sporozoites from oocysts, which infects vertebrate hosts very poorly (Touray et al. 1992). Invasion of the salivary glands could trigger certain developmental changes in the sporozoites that result in higher infectivity (Matuschewski et al 2002b, Kaiser et al. 2004). Upon successful infection of the hepatocyte, the sporozoite transforms and grows and multiplies forming the liver schizont. After a few days thousands of merozoites are released into the blood stream. Upon reaching the blood stream, they invade the red blood cells and start the asexual cycle. Currently, only a few sporozoites proteins have been identified that play a role in establishing infection of the liver. CSP and TRAP in addition to their role in salivary gland invasion are also required for invasion of hepatocytes (Cerami et al. 1992, Matuschewski et al. 2002a). CSP has been shown to interact with glycosaminoglycans present in the liver (Pinzon-Ortiz et al. 2001, Ancsin and Kisilevsky 2004). CSP also undergoes proteolytic processing and cysteine protease inhibitors block this processing (Coppi et al. 2005) as a result inhibiting liver invasion. Animals immunized against CSP are completely protected against sporozoite challenge (Bruna-Romero et al. 2001). 30

33 TRAP on the other hand is secreted during invasion of the hepatocyte. The sporozoite can enter hepatocytes either by disruption of the membrane or by invagination of the membrane followed by membrane closing. In the first pathway, the sporozoite enters and exits the cell, while the second pathway leads to a productive infection through formation of a parasitophorous vacuole that surrounds the parasite. Recognition and invasion of a hepatocyte is a complex process involving traversal through the liver sinusoidal layer that contains the macrophage-like Kupffer cells (Mota et al. 2002a) and subsequently through several hepatocytes (Mota et al. 2002a, Mota and Rodriguez 2004) before committing to the transformation process. This cell-traversal induces the release of hepatocyte growth factor (HGF). HGF in turn primes the hepatocytes for infection by activating the HGF signaling cascade (Carrolo et al. 2003, Leiriao et al. 2005). Sporozoites that go through the enter and exit process have an accumulation of TRAP at their apical end (Mota et al. 2002). This accumulation of TRAP might result in interaction with certain receptor(s) on the hepatocyte surface that signal the parasite for a productive infection. Antibodies against TRAP does not inhibit sporozoite invasion in a rodent malaria model (Gantt et al. 2000). As most of the protein is surface exposed immediately prior to invasion, the antibodies might not have enough access to inhibit invasion. TRAP appears to be trafficked by two different secretory processes, one that presumably bypasses the micronemes and is transported to the surface of sporozoites and the other that is trafficked through the micronemes and is surface exposed only during the contact of the activated 31

34 sporozoite with the hepatocyte. The later probably has a function in invasion and the former is required for sporozoite motility and probably is also involved in invasion. What regulates the two different secretory processes is not known. PbCellTOS is a micronemal protein expressed in both ookinetes and sporozoites (Kariu et al. 2006). Targeted disruption of the gene resulted in a severe impairment of both the ookinete invasion of the mosquito midgut as well as sporozoite infection of the liver. PbCellTOS(-) sporozoites were defective in infecting mice, however they readily infected Kuffer cell-depleted mice. CellTOS seems to be required for the sporozoites to enter through Kuffer cells in order to invade the heaptocytes. This also indicates that there are conserved cell passage mechanisms that the parasite uses to invade different host cell barriers. PbSPECT, a sporozoite microneme protein essential for cell traversal, belongs to the group of perforin domain-containing proteins (Ishino et al. 2005). The other member of this group is expressed in ookinetes (see above). The sporozoites have to first enter the sinusoidal layer before they reach the hepatocytes. It is speculated that the sporozoites use SPECT to pass through the sinusoidal layer by forming pores on the memebrane. In the absence of this protein, parasites are no longer are able to pass thorough the sinusoidal layer and remain in circulation (Ishino et al. 2005). As mentioned earlier, sporozoites go through a enter and exit process before committing to productive infection. The factors that are involved in this process are 32

35 not clearly understood. Two recent articles indicate that novel proteins, P36p and P36, of the 6-Cys domain protein family (6 conserved cysteine residues at precise distance in the protein), are required for the sporozoites to commit to productive infection (Ishino et al and van Dijk et al. 2005). Parasites lacking P36p and P36 protein were able to invade hepatocytes in vitro, but could not commit to infection. In addition P36p infected hepatocytes undergo enhanced apoptosis (van Dijk et al. 2005). Mice immunized with P36p(-) sporozoites are protected against a subsequent wild type challenge (van Dijk et al. 2005). Two genes that are up-regulated in sporozoties only after salivary gland invasion (UIS3 and UIS4) were shown to be required for parasite development inside the hepatocytes (Mueller et al. 2005a,b). They showed that parasites lacking these two proteins invaded and started development inside hepatocytes but were soon arrested. Further these UIS3(-) and UIS4(-) sporozoites could be used as genetically attenuated vaccines. Mice immunized with these sporozoites were protected upon challenge with wild type sporozoites (Mueller et al. 2005a,b). Study of gene function in Plasmodium parasites. The genome of the parasite is haploid during its development in the vertebrate host. Though the DNA content of the gametocytes increases, it is only after fertilization that they become diploid. However, they return back to haploidy soon after sporozoite formation. Study of gene function in Plasmodium takes advantage of this feature, whereby there is no need for the time consuming crosses to generate homozygosity. Gene disruption is achieved by 33

36 transfecting either circular (P. falciparum) or linear (P. berghei and P.yoelli) constructs carrying the homologous regions for recombination into asexual stage parasites (de Koning-Ward et al. 2000). This technique provides a unique opportunity to understand the complex process of Plasmodium development. The first demonstration of this technique came in 1997, which elucidated the role of a major sporozoite surface protein, circumsporozoite protein (CSP) in sporozoite development (Menard et al. 1997). This technique allows the study of function of those proteins that are required for development in the mosquito, as disruption of such genes will have no adverse effect on the blood stage development. However genes that are essential for asexual stage development cannot be disrupted by this method. To study such genes, a new strategy of conditional mutagenesis has recently been developed (Nkrumah et al. 2006). 34

37 Research Objectives Malaria is a devastating global disease that afflicts more that 10% of the world s population. This disease claims between 1-3 million lives every year, mainly children under the age of five. Plasmodium, the etiological agent of malaria is an obligate intracellular parasite. Disease transmission begins with the bite of an infected female Anopheles mosquito. Sporozoites, the infective form of the parasite, present in the salivary glands of infected mosquitoes are injected into the vertebrate host during blood feeding. These parasites infect and develop in the hepatocytes before releasing thousands of merozoites, the disease causing forms. The merozoites invade and develop within RBCs and produce more merozoites and this cycle continues (asexual cycle). A small proportion of these parasites commit to sexual differentiation. Sexual development begins when a mosquito takes up the infected blood. Parasite development in the mosquito is a highly complex process that involves the formation of six completely different forms of the parasite. The motile ookinetes formed as a result of fertilization, actively invade the midgut epithelium and transform into a round oocyst. Midgut invasion involves interaction of parasite surface proteins with midgut receptors. Invasion triggers a potent mosquito immune response, killing a large number of parasites. However, those few ookinetes that are able to survive this onslaught emerge on the basal side of the midgut and transforms into non-motile oocysts. The oocyst is surrounded by a thick capsule and appears to be resistant to the mosquito defenses. It is also possible that the oocyst fails to trigger an immune response. Each oocyst produces thousands of sporozoites, which invade 35

38 the salivary glands and are now ready to be transmitted. The molecular mechanisms that regulate some of these processes are poorly understood. The main aim of the present study is to identify some of the players involved in oocyst development and sporozoite formation. In doing so, this study also sheds light on another important feature, namely, host-cell invasion. As a first step, we set out to identify genes that are up-regulated during oocyst development. The oocyst stage of the parasite is intricately associated with the midgut and basal lamina of the mosquito. Hence, it is not possible to isolate oocysts from the mosquito tissue for molecular studies. We used a subtractive hybridization approach to enrich for genes that are specifically expressed during oocyst development in the mosquito. This approach will also identify mosquito genes that are expressed in response to the developing oocyst. We identified several novel parasite genes that are induced at different times during oocyst development. In addition this study also identified parasite surface proteins that were initially thought to be expressed only in blood-stage parasites. Genes that are involved in oocyst development and differentiation would be expected to be significantly up-regulated in the oocyst. Using this criterion we identified PbCap380, a gene that is up-regulated only during oocyst development. We speculated that PbCap380 plays an important role in parasite development in the mosquito. To test this hypothesis, we used immunological and genetic techniques to dissect the function of the protein. Oocyst capsule is speculated to protect the 36

39 developing parasite from the mosquito immune system. Our data shows that interfering with a single component of the oocyst capsule results in the parasite being killed by the mosquito. This study identifies the parasite oocyst as a potential target for blocking malaria parasite transmission. We also identified a rhomboid-family serine protease (ROM1) in the subtraction library. Recent studies in P. falciparum identified ROM1 to be expressed also in the asexual stage parasites. Homology searches identified a second rhomboid protein, ROM3, to be expressed in sexual stage parasites. In Drosophila, rhomboid proteins regulate epidermal growth factor receptor (EGFR) signaling by intramembrane processing of the EGF ligand Spitz. This signaling regulates many aspects of growth and differentiation. We hypothesized that ROM1 and ROM3 are required for parasite development. Genetic studies reveal distinct roles for ROM1 and ROM3 at multiple stages of parasite invasion, development and disease pathogenesis. 37

40 CHAPTER 2 Analysis of Plasmodium and Anopheles Transcriptomes during Oocyst Differentiation Prakash Srinivasan 1, Eappen G. Abraham 1, Anil K. Ghosh 1, Jesus Valenzuela 2, Jose M. C. Ribeiro 2, George Dimopoulos 3, Fotis C. Kafatos 4, John Adams 5, Hisashi Fujioka 4 and Marcelo Jacobs-Lorena 1 1 Department of Genetics, Case Western Reserve University, Cleveland, OH 44106; 2 Medical Entomology Section, Laboratory of Parasitic Diseases, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD ; 3 Department of Biological Sciences, Centre for Molecular Microbiology and Infection, Imperial College, London SW72AZ, United Kingdom; 4 European Molecular Biology Laboratory, Heidelberg, Germany; 5 Department of Biological Sciences, University of Notre Dame, Notre Dame, IN 46556; 4 Department of Pathology, Case Western Reserve University, Cleveland, OH Present address: Dept. Molecular Microbiology & Immunology and Malaria Research Institute, Johns Hopkins School of Public Health, Baltimore, MD Reference: Srinivasan et al, J Biol Chem

41 Abstract Understanding the life cycle of the malaria parasite in its mosquito vector is essential for developing new strategies to combat this disease. Subtractive hybridization cdna libraries were constructed that are enriched for Plasmodium berghei and Anopheles stephensi genes expressed during oocyst differentiation on the midgut. Sequencing of 1485 random clones led to the identification of 1137 unique expressed sequence tags. Of the 608 expressed sequence tags with data base hits, 320 (53%) had significant matches to the non-redundant protein database, whereas 288 (47%) with matches only to genomic data bases represent novel Plasmodium and Anopheles genes. Transcription of six novel parasite genes and two previously identified asexual stage genes was up-regulated during oocyst differentiation. In addition, the mrna for an Anopheles fibrinogen domain gene was induced on day 2 after an infectious blood meal, at the time of ookinete to oocyst differentiation. The subcellular distribution of MAEBL, a sporozoite surface protein, is developmentally regulated from presumed storage organelles in day 15 oocysts to uniform distribution on the surface in day 22 oocysts. This redistribution may reflect a sporozoite maturation program in preparation for salivary gland invasion. Furthermore, apical membrane antigen 1, another parasite surface molecule, is translationally regulated late in sporozoite development, suggesting a role during infection of the vertebrate host. The present results and those of an accompanying report (Abraham et al. 2003) provide the foundation for studies seeking to understand at the molecular level Plasmodium development and its interactions with the mosquito. 39

42 Introduction Malaria is the major vector-borne infectious disease. Nearly 40% of the world's population is at risk, and close to 2 million persons (mostly children under the age of five) die every year. Despite considerable efforts to contain the disease, the number of infected persons is raising (Breman et al. 2001). The lack of an effective vaccine and the emergence of drug-resistant parasites and insecticideresistant mosquitoes warrant the search for alternate strategies to combat this disease. The parasite undergoes a complex developmental program in the mosquito (Ghosh et al. 2000, Sinden, 1998). After traversing the midgut, the parasite lodges itself between the midgut epithelium and the basal lamina, giving rise to an oocyst. Upon maturation, sporozoites formed inside the oocyst are first released into the hemocoel and then invade the salivary glands. To date only a handful of parasite genes expressed specifically during differentiation in the mosquito have been identified. Oocyst differentiation, a process that leads to sporozoite formation, takes close to 2 weeks. Identifying genes that are required for this process and for the subsequent release of sporozoites could provide novel targets to interfere with parasite development. Thus far only two genes, namely circumsporozoite protein (CS) 1 and a scavenger receptor-like protein, have been shown to be required for sporozoite formation (Menard et al. 1997, Claudianos et al. 2002). The mosquito responds to the presence of the parasite by mounting an immune response. A variety of immune genes are induced during ookinete penetration of the midgut and again at the time of 40

43 sporozoite release into the hemocoel (Dimopoulos et al. 1997). The parasite elicitors of these responses are not known. The lack of methods to fractionate oocysts from mosquito midgut tissues has hindered the molecular analysis of oocyst development. An additional limitation is that only a small proportion of the RNA extracted from infected mosquito guts originates from the parasite. In an attempt to overcome some of these limitations we used a subtractive hybridization strategy to enrich for Plasmodium berghei and Anopheles stephensi genes that are induced during oocyst development in the midgut. Numerous novel genes both of parasite and of mosquito origin were identified. A mosquito fibrinogen-like gene that is induced by the parasite and two Plasmodium surface proteins, MAEBL and apical membrane antigen 1 (AMA1), were further characterized. Materials and Methods Parasites maintenance. P. berghei ANKA strain, clone 2.34, was maintained by passage in Swiss Webster mice. Mice with 15-20% parasitemia and 2-3 exflagellations per field (40x magnification) were used for mosquito infections. Mosquito Infection. A. stephensi mosquitoes were fed on infected mice in a 21 C incubator. Mosquito infection was assessed by counting oocyst numbers of guts dissected on day 15 after an infectious blood meal. For each experiment the average oocyst number per gut was 75 or higher. 41

44 RNA Preparation and Subtraction Library Construction. Total RNA was isolated with Trizol (Molecular Research Center) and polyadenylated RNA was prepared by chromatography on an oligo(dt) column. Polyadenylated RNA corresponding to 75 µg of total RNA of each sample was treated with DNase I (Invitrogen) to remove any remaining genomic DNA. The samples were then reverse-transcribed using the SMART cdna synthesis primer (modified oligo(dt), Clontech), as described by the manufacturer. Equal amounts of cdna from each sample was pooled as indicated in Fig.2.1 and subtraction was carried out using the Clontech PCR Select cdna subtraction kit (catalog number K1804-1). The subtraction products were cloned into pgem-t easy vector (Promega) and transformed into Escherichia coli (DH5 ) competent cells. Inserts were PCR amplified using T7 (5'- CTAATACGACTCACTATAGGGC-3') and SP6 (5'- GATTTAGGTGACACTATAG-3') primers and sequenced with a nested primer, PN1 (5'-TCGAGCGGCCGCCCGGGCAGGT-3') in an ABI377 sequencer. Clustering and Data Base Analysis. Vector sequences were first removed computationally from all sequences followed by clustering of identical clones by internal BLAST analysis. Each cluster was searched for homology using the nonredundant data base and the Malaria Genetics and Genomics data bases at the National Center for Biotechnology Information (NCBI) using the BLASTN and BLASTX algorithm. In addition, sequences were also searched against the Plasmodium data base at The Institute of Genomic Research ( and Sanger ( Sequences with significant BLASTX similarity (P(N) 10-5 ) were grouped based on the predicted function of the homologous protein. 42

45 Sequences with no significant BLASTX similarity were grouped based on their BLASTN similarity (P(N) ). Preparation of cdna Blots. After PCR amplification, subtracted and non-subtracted cdnas were fractionated by electrophoresis on 1% agarose gels and transferred onto positively charged nylon membranes. A mosquito actin probe was generated by PCR with primers actin F (5'-TCAAGATGTGCGACGAAGAGG-3') and actin R (5'- CGAGAGATGTGGTGTGTTGTTTCTC-3') and labeling with [P 32 ]dctp using random hexamers. Reverse Transcription-PCR Expression Analysis. RNA was isolated from midguts dissected at various times after an infectious or a non-infected blood meal and from blood-stage parasites. Total RNA (1 µg) from each sample was reverse-transcribed with an oligo(dt) primer and PCR amplified (25 cycles) using the following genespecific primers: PbRP.F (5'-GACTAACAAGAGCGGCAAGA-3') and PbRP.R (5'- GTACATAAAATCCCATTCCATAT-3'); CS.F (5'- GTACCATTTTAGTTGTAGCGTC-3') and CS.R (5'- CATCGGCAAGTAATCTGTTG-3'); AMA1.F (5'-CCTTCAGGTAAATGTCCAGT- 3') and AMA1.R (5'-TTTCCCAATCATCACGCA-3'); M184.F (5'- GCCTAAATGTGTTGCGAAGAGC-3') and M184.R (5'- CAGAAAGGGTGTATAATGATGG-3'); L1144.F (5'- AAGGATGGTTTTGACTGTTT-3') and L1144.R (5'- ATTTAATTTCTCCAATGGGT-3'); L1147.F (5'-TAAAAGGCAGCAATAACAG- 3') and L1147.R (5'-CATCAACAGATCCGACAC-3'); L457.F (5'- 43

46 ATGCCTCTTTTATCTGTCTA-3') and L457.R (5'- GATTACCAGCCGTATGTTGC-3'); L1085.F (5'-TATCATATTCCGCCTTTGC-3') and L1085.R (5'-TGTGTGCATATTTACATAATCCT-3'); L1096.F (5'- TGATGAAACAATCCAAACTG-3') and L1096.R (5'- AATTTATCATGGCCCAACT-3'); L2451.F (5'-TGGAGCAACTTGGTGTAA-3') and L2451.R 5'-CCAGTATTGGTAGGGAAT-3'); L3593.F 5'- ATTCCGAAGTTTAATACCAA-3') and L3593.R (5'-TATGACCAGCTTTAGCAC- 3'). Primers were designed such that their annealing temperature was C and the product size bp. The PCR products were blotted and hybridized with a radioactive probe prepared from DNA amplified with the same primer using the cloned insert as the template. Indirect Immunofluorescence. Guts were dissected on days 15 and 22, and salivary glands were dissected on day 25 after an infectious blood meal. Sporozoites obtained by gentle homogenization of guts or salivary glands were spotted on glass slides and fixed in ice-cold methanol. For double labeling, the fixed sporozoites were then incubated for 2 h in 5% bovine serum albumin in phosphate-buffered saline followed by incubation for 1 h with polyclonal antisera against the carboxyl cysteine-rich region of P. berghei MAEBL and a monoclonal antibody against CS (3D11, kindly provided by Dr. Victor Nussenzweig). Monoclonal antibody 28G2dc1 (kindly provided by Dr. Alan Thomas) that recognizes a highly conserved C-terminal region from several Plasmodium species was used to localize AMA1. The following secondary antibodies were used: rhodamine-conjugated goat anti-rabbit IgG (Sigma), 44

47 fluorescein isothiocyanate-conjugated rabbit anti-mouse IgG (Sigma), and rhodamineconjugated rabbit anti-mouse IgG (Sigma). Immuno-electron Microscopy. P. berghei sporozoites isolated from salivary glands were fixed for 20 min at 4 C with 1% formaldehyde and 0.2% glutaraldehyde in 0.1 M phosphate buffer, ph 7.4. Fixed samples were washed, dehydrated, and embedded in LR White resin (Polysciences). Thin sections were blocked for 30 min at room temperature in phosphate-buffered saline containing 5% w/v nonfat dry milk and 0.01% v/v Tween 20 (PBTM). Grids were then incubated with anti-ama 1 primary antibody (diluted 1:50 in PBTM for 16 h at 4 C. Control grids were incubated with normal mouse serum in PBTM at the same dilution. After washing grids were incubated for 1 h in 15-nm gold-conjugated goat anti-mouse IgG (Amersham Biosciences) diluted 1:20 in phosphate-buffered saline containing 1% w/v bovine serum albumin and 0.01% v/v Tween 20 (PBTB), rinsed with PBTB, and fixed with 2.5% glutaraldehyde to stabilize the gold particles. Samples were stained with uranyl acetate and lead citrate and then examined in a Zeiss CEM902 electron microscope (Oberkochen, Germany). Rapid Amplification of cdna Ends. Full-length cdna for A. stephensi fibrinogen domain protein 1 was isolated by rapid amplification of cdna ends using primers M2015F (5'-GTTCGAAATTGGAGACGAGC-3') and M2015R (5'- CACCATGCTCCAGCAAACG-3'). Sequence analysis was performed at the Expasy Molecular Biology Server ( Sequence alignment was done using ClustalW software. 45

48 Results Construction and Characterization of the Subtraction Libraries. To construct libraries enriched for P. berghei and A. stephensi transcripts expressed during midoocyst (day 4-8) and late-oocyst (day 10-14) stages of parasite development, gut cdnas of non-infected mosquitoes and of blood-stage parasites were subtracted from the cdnas of guts harboring developing oocysts (Fig.2.1). This procedure was expected to enrich for parasite transcripts expressed during oocyst differentiation and for mosquito transcripts induced in the gut by the parasite. Any common transcripts (e.g. genes expressed in both infected and non-infected guts or genes expressed in both blood-stage and mosquito-stage parasites) are depleted by the procedure. Subtraction efficiency was assessed by hybridizing a mosquito actin probe to Southern blots of cdnas obtained before and after the subtraction procedure. The signal was much weaker in the "after subtraction" lanes, indicating effective depletion of common genes by the subtraction protocol (data not shown). A total of 1485 clones from both libraries was sequenced and analyzed (Table2.1). The 804 sequences from the mid-oocyst library could be grouped into 689 unique clusters, of which 623 were singletons (sequences isolated only once). Of the 689 sequences, 367 (53%) had data base hits and were grouped according to their presumed function (Fig.2.2A). Similarly, 681 sequences from the late-oocyst library could be grouped into 448 unique clusters, of which 334 were singletons. Of the 448 sequences, 241 (55%) had data base hits and were classified based on their predicted function (Fig.2.2B). 46

49 Fig.2.1. Construction of subtraction libraries. For each library a mixture of equal amounts of cdna from each of the samples listed in the middle set of three boxes was subtracted from a pool of equal amounts of cdna from samples in the Infected midgut box at the top. Infected blood, cdna from blood-stage parasites obtained from infected mice; Uninfected and infected midgut, cdna prepared from mosquito midguts dissected at the indicated times (hours or days) after a non-infected or infected blood meal, respectively. Unfed, cdna from midguts of mosquitoes that did not have a blood meal. The high proportion of singletons in both libraries suggests that the subtraction procedure resulted in efficient normalization (removal of redundant sequences). As construction of the library involves digestion of cdnas with a restriction enzyme, some of the clusters might represent non-overlapping regions of the same gene. A number of mosquito genes involved in maintenance of the redox state of the cell were identified. These include ESTs with homology to thioredoxin, peroxiredoxin-1, metallothionein, glutathione transferase, and glutathione peroxidase (Fig. 2.2 A and 47

50 B). In addition, several components of the electron transport chain, which is a major source of reactive oxygen species, were also identified. Interestingly, transcripts from putative anti-apoptotic genes such as inhibitor of apoptosis 1 (IAP1) and Bax inhibitor 1 were found in mid- and late-oocyst libraries, respectively. This is in contrast to the ookinete/early oocyst library, which contained ESTs corresponding to apoptosispromoting rather than anti-apoptotic genes (Abraham et al. 2002). Table 2.1. Characteristics of the subtraction libraries. Singletons refer to ESTs found only once. Clusters with database similarities were based on BLASTN and/or BLASTX searches at NCBI, PlasmoDB, TIGR, and Sanger with a probability cutoff of P (N) e 10 (BLASTN) or P (N) e 5 (BLASTX). 48

51 Fig.2.2. Functional classification of ESTs based on BLASTX and BLASTN similarities. A total of 363 unique sequences from the mid-oocyst library and 245 unique sequences from the late-oocyst library had data base hits (Table I) and were grouped based on their predicted biological function. A, mid-oocyst library. B, late-oocyst library. Genomic Survey Sequence (GSS) is a sequence that has similarity with genomic DNA. 49

52 A. stephensi Fibrinogen Domain Protein 1 (AsFBN1) Is Induced by the Parasite. Two overlapping ESTs (GenBank accession numbers CB and CB602444) with similarity to fibrinogen domain proteins were isolated from the mid-oocyst library. There is evidence that in mosquitoes and horseshoe crabs, members of this gene family are induced in response to foreign organisms (Dimopoulos et al. 2002, Gokudan et al. 1999). The full-length cdna was cloned by 5'- and 3'-rapid amplification of cdna ends. The resulting 921-bp sequence predicts a protein of 306 amino acids, including a 19-amino acid secretion signal sequence at the N terminus and a 200-amino acid fibrinogen domain at the C terminus (Fig.2.3A). AsFBN1 is related to horseshoe crab tachylectins 5A and 5B (25% identity, 26% similarity), suggesting that it could act as a broad specificity immune molecule by binding carbohydrate molecules on glycoproteins (Gokudan et al. 1999). AsFBN1 and the Anopheles gambiae homologue are 72% identical, including six conserved cysteine residues in the secreted portion of the protein and two in the signal peptide (Fig.2.3A). Analysis of AsFBN1 gene expression revealed that it is induced at 48 h by a parasitecontaining blood meal but not by a non-infected blood meal (Fig.2.3B and C). This suggests that induction is likely to be triggered by the presence of the differentiating oocyst. Temporal Patterns of Plasmodium Gene Expression. Reverse transcription-pcr was used to assess changes in parasite mrna abundance as a function of time after mosquito infection (Fig2.4A). In initial experiments, the assay was run by resolving PCR products on agarose gels followed by ethidium bromide staining. 50

53 Fig Analysis of the AsFBN1. A, sequence comparison of the A. stephensi (As) and A. gambiae (Ag) fibrinogen domain protein genes with horseshoe crab tachylectins 5A and 5B. Amino acid residues that are conserved between all the four proteins are shaded light, and residues that conserved within the mosquito proteins are shaded dark. Arrow, putative signal peptide cleavage site; Dotted line, fibrinogen-like domain; *, conserved cysteine residues. AgFBN24, A. gambiae fibrinogen domain protein 24; TL5A, tachylectin 5A; TL5B, tachylectin 5B. B, AsFBN1 is induced by Plasmodium in the mosquito midgut. Northern blot analysis shows that the gene is induced at 48 h after an infected but not after a non-infected, blood meal. The numbers refer to the time in hours after a blood meal at which the midguts were dissected. 0h, sugar-fed mosquitoes; UInf, uninfected blood meal; Inf, infected blood meal. C, Northern blots such as the one shown in B were quantified using phosphorimaging. The signals were normalized to a mosquito loading control (mitochondrial rrna) and are plotted relative to the value of sugar-fed controls. Gray and black bars represent expression profiles from two independent sets of RNA samples. 51

54 Quantification by this protocol was not reliable because many of the target mrnas were rare in the samples and required large number of PCR amplification cycles for detection. The radioactive hybridization approach (cf. "Materials and Methods") proved to be much more sensitive, linear, and reproducible. Fig.2.4B shows that under the conditions used the signal is proportional to RNA concentration. To partially compensate for the presence of large numbers of blood-stage parasites in the 24h inf sample, more template was used for this than later time points (Fig.2.4A PbRP). Note that all blood-stage parasites eventually die and are eliminated from the mosquito guts. CS, which is expressed during oocyst development (Menard et al. 1997), was used as a positive control (Fig.2.4A). To analyze temporal expression patterns, eight ESTs that had hits to the Plasmodium data bases were chosen (Fig.2.4A). Of these, six (L1085, L2451, L1147, L1144, M184, L1096) are novel, and two (L457 and L3593) had been isolated previously. All but L1096 and L3593 appear to be induced specifically during the mosquito stages and have an expression profile consistent with the library (mid-oocyst or late-oocyst) from which they were derived (Fig.2.4A). L1085, L2451, and L1147 start being expressed at different times but all have peak expression during the terminal stages of oocyst differentiation (day 15). These gene products could function in final stages of oocyst differentiation and/or be stored for use during sporozoite invasion of the mosquito salivary gland or vertebrate liver. L2451 encodes a thioredoxin domain protein. Thioredoxins play a crucial role in defense against free radicals (Rahlfs et al. 2002, Aslund and Beckwith 1999) and may be involved in protection against reactive 52

55 oxygen species generated by the mosquito immune system. L1147 may also be expressed in ookinetes (24 h). Fig.2.4. Temporal expression patterns of selected ESTs from the subtraction libraries. A, reverse transcription-pcr products of samples indicated above each lane were amplified for 25 cycles with primers for genes specified to the left of each panel, fractionated by gel electrophoresis, and transferred onto nylon membranes. Radioactive probes generated by PCR amplification of the corresponding cloned EST were hybridized to the membranes. Plasmodium ribosomal protein gene (PbRP; accession number BF295783) was used as a 53

56 loading control, and the circumsporozoite protein gene (accession number M14135) was used as a positive control. Similar profiles were obtained in at least four experiments with at least two independently isolated RNA samples. M, clones from the mid-oocyst library; L, clones from the lateoocyst library; bl, RNA from blood stage parasites; 24 h N-in, RNA from guts dissected 24 h after a non-infected blood meal; 24 h Inf, RNA from guts dissected 24 h after an infected blood meal; D4, D9, and D15, RNA from guts dissected at the indicated number of days after an infected blood meal; Unknown, EST having homology to the Plasmodium genomic data base that does not have a predicted open reading frame; Hypothetical, EST having homology to the Plasmodium genomic data base that has a predicted open reading frame. B, linearity of signal response. Increasing amounts of template (relative amounts indicated at the top of each lane) were amplified for 25 cycles and analyzed as in panel A. Note that the strength of the signal is proportional to the amount of RNA template. M184 and L1144 had peak expression at mid-oocyst stages (day 9) followed by a sharp decrease of transcript abundance, indicating that these gene products may be required for intermediate steps in sporozoite differentiation. L1096 is expressed at all stages of parasite development. The predicted protein has homology to halo-acid dehalogenase family proteins. This family includes L-2-haloacid dehalogenase (Tsang and Sam 1999) and epoxide hydrolases (Argiriadi et al. 1999) that function in halocarbon metabolism and detoxification of epoxide substrates, respectively. This family also includes certain phosphatases. The expression of two ESTs (L457 and L3593) with similarity to previously identified Plasmodium genes were also analyzed (Fig.2.4A). L3593 encodes a homologue of Rab6, a protein associated with the Golgi apparatus of blood-stage P. 54

57 falciparum (Van Wye et al. 1996). This is of interest because there has been no morphological evidence of a functional secretory pathway in the mosquito stages of parasite development. The finding that L3593 is expressed during sporozoite differentiation raises the possibility that the oocyst employs a Golgi-like pathway for protein trafficking. MAEBL (L457) and AMA1, two proteins originally thought to be expressed only in blood stages, were found to be strongly activated at late stages of oocyst differentiation (Fig.2.4A). Further studies on these two genes are described below. Differential Localization of MAEBL during Sporozoite Maturation. Two nonoverlapping ESTs (GenBank accession numbers CB and CB603325) corresponding to the C-terminal region of the erythrocyte-binding protein MAEBL were recovered from the late-oocyst library. MAEBL was initially identified as a blood-stage gene whose protein localizes to the apical organelles of merozoites (Kappe et al 1997, Kappe et al. 1998). Recently, MAEBL was reported to be expressed also in sporozoites (Kappe et al. 2001). We found that the gene is sharply induced during late stages of oocyst maturation (Fig.2.4A). Indirect immunofluorescence revealed that MAEBL subcellular localization changes as the sporozoite matures. In immature sporozoites (day 15) the protein was restricted to the apical end, presumably confined to secretory organelles (Fig.2.5). In contrast, CS, which is another major sporozoite surface protein, was uniformly distributed in the same sporozoites. In mature midgut sporozoites (day 22) and in sporozoites that had invaded the salivary glands both MAEBL and CS were uniformly distributed on the sporozoite surface. The deployment of the protein to the sporozoite surface upon 55

58 maturation is consistent with the observation that MAEBL is essential for salivary gland invasion (Kariu et al. 2002). The mechanism that regulates MAEBL translocation from the apical end to the surface is unknown. In blood stage parasites, MAEBL has been shown to transit through the ER-Golgi secretory pathway. The sporozoites could employ a similar secretory system. Reverse transcription-pcr performed using primers representing both N- and C-terminal regions of the coding sequence indicates that MAEBL is also expressed in liver-stage parasites (results not shown). Thus, MAEBL may be required both for salivary gland and liver invasion, as is the case for CS (Warburg et al. 1992, Cerami et al. 1992). Fig.2.5. Differential localization of P. berghei MAEBL in oocyst and salivary gland sporozoites. Anti-MAEBL and anti-cs antibodies were used to localize the corresponding proteins on sporozoites. The source of sporozoites was as follows: MG D15, midguts on day 15 after infection; MG D22, midguts on day 22 after infection; SG D25, salivary glands on day 25 after infection. The sporozoites were counterstained with 4,6-diamidino-2-phenylindole (DAPI) to show the location of the nucleus. The MG D15 pattern was consistently observed in three independent preparations. No reproducible differences of protein distribution between day 22 MG and day 25 SG sporozoites were observed. 56

59 B Fig.2.6. Regulation of AMA1 expression in sporozoites. A, anti-ama1 and anti-cs antibodies were used to localize the corresponding proteins on sporozoites. The source of sporozoites was as follows: MG D15, midguts on day 15 after infection; MG D22, midguts on day 22 after infection; SG D25, salivary glands on day 25 after infection. Sporozoites with no AMA1 expression were stained with 4,6-diamidino-2-phenylindole, which labels the nucleus, to show the presence of the parasite. B, cross-sectional view of salivary gland sporozoites labeled with anti-ama1 antibody. Immuno-electron microscopy shows the surface localization of AMA1 on salivary gland-invaded sporozoites. No staining was observed with midgut sporozoites (data not shown). AMA1 Expression Is Post-transcriptionally Regulated during Oocyst Development. The expression of MAEBL during oocyst development prompted the analysis of other blood-stage parasite surface proteins, including MSP1 and AMA1. 57

60 MSP1 expression seems to be restricted to blood forms of the parasite (not shown). However, AMA1 is strongly induced during the final stages of oocyst maturation (Fig.2.4A). In contrast to CS and MAEBL, AMA1 protein expression could not be detected in sporozoites isolated from either day 15 or day 22 oocysts (Fig.2.6A). Interestingly, protein can only be detected after sporozoite invasion of salivary glands (day 25, Fig.2.6A). Moreover, immuno-electron microscopy revealed that AMA1 is localized on the surface of salivary gland sporozoites (Fig.2.6B). In agreement with the immunofluorescence results (Fig.2.6A) no immunogold staining was observed with midgut sporozoites (not shown). Thus, the presence of the transcript (Fig.2.4A) but not of the protein (Fig.2.6A) in midgut sporozoites indicates that AMA1 is translationally regulated during sporozoite maturation. The presence of the AMA1 protein only on the surface of salivary gland sporozoites (Fig.2.6B) suggests that AMA1 might be required for invasion of hepatocytes. Discussion In this study we report the identification of candidate genes of both mosquito and parasite origin that are induced during oocyst development on the mosquito midgut. The lack of suitable techniques to purify oocysts from mosquito midgut tissues and the small proportion of parasite RNA in total RNA from infected midguts have seriously hampered the molecular analysis of Plasmodium development in the mosquito. The recent demonstration of sexual stage parasite development in an in vitro system (Al-Olayan et al. 2002) has yet to be adapted for large-scale cultures. 58

61 The present study has partially overcome these limitations by the construction of stage-specific subtraction libraries. By reverse transcription-pcr we demonstrate that these libraries indeed contain genes induced during oocyst development on the mosquito midgut. Ongoing microarray analysis of clones from the two subtraction libraries promises to provide a more detailed survey of mosquito and parasite genes differentially regulated during oocyst differentiation. A large proportion of the Plasmodium life cycle in the mosquito is dedicated to the differentiation of the oocyst, resulting in the formation of thousands of sporozoites. Parasite genes expressed at specific stages of the differentiation process are likely to participate in these events. For instance, M184 and L1144 may be required for intermediate steps in sporozoite differentiation because they have peak expression at midoocyst stages (day 9, Fig.2.4A) and L1085, L2451, and L1147 may be required in terminal stages, as indicated by their activation late during oocyst differentiation (day 15, Fig.2.4A). It is well established that the mosquito responds to the presence of the invading parasite (ookinete) by mounting an innate immune response (Dimopolous et al. 1997). Less well known are the responses against the developing oocyst. This study provides the first insights on the response of the mosquito to this stage of the parasite. Some of the transcripts identified include genes with homology to antioxidants and key antiapoptotic proteins such as inhibitor of apoptosis 1 and Bax inhibitor 1. Annotation of the recently completed A. gambiae genome revealed the presence of a large family of fibrinogen-related proteins (FBN) (Christophides et al. 2002, Zdobnov et al. 2002). Some FBN family genes are induced either immediately 59

62 after a bacterial challenge or 24 h after a Plasmodium-infected blood meal (Dimopoulos et al, 2002). In contrast, As- FBN1 is induced on day 2 after an infectious blood meal, at which time ookinete-to-oocyst transformation is completed. It is likely that this induction is triggered by the developing oocyst. Because fibrinogen can interact with thrombospondin I (Bonnefoy ey al. 2001), the possibility arises that AsFBN1 recognizes thrombospondin I domains of Plasmodium surface proteins, such as circumsporozoite protein and thrombospondin-related adhesive protein. Surprisingly, MAEBL localization was found to be developmentally regulated in sporozoites, from apical localization (presumably in secretory organelles) in immature sporozoites to uniform distribution on the entire surface in mature sporozoites. After invasion of the salivary glands, MAEBL remains on the sporozoite surface, indicating that the protein might be required for vertebrate stage (liver) infection. The mechanism that regulates MAEBL translocation from the apical end to the surface is unknown. AMA1 has been shown in various Plasmodium species to be required for invasion of erythrocytes (Triglia et al. 2000) and as such is a candidate for a bloodstage vaccine in both rodent and primate models (Narum et al. 2000, Stowers et al. 2002). AMA1, which was thought to be a surface protein specific to the asexual stages of the parasite life cycle, was found to be strongly induced in day 15 oocysts. The mrna is translationally regulated, since the protein appears on the sporozoite surface only after invasion of the salivary gland. 60

63 This raises the intriguing possibility that AMA1 is required during invasion of the vertebrate liver. The ESTs generated in this study will contribute to the annotation of the recently completed Plasmodium yoelii and A. gambiae genomes (Gardner et al. 2002, Holt et al. 2002). For instance L1085 has homology to the Plasmodium nucleotide database but has no predicted open reading frame. Many of the genes identified in this study are novel, and elucidation of their function will provide a stronger foundation for understanding parasite development and Plasmodium interactions with the mosquito. 61

64 CHAPTER 3 PbCap380, a novel Plasmodium oocyst Capsule Protein is Essential for Parasite Survival in the Mosquito Prakash Srinivasan, Hisashi Fujioka and Marcelo Jacobs-Lorena Reference: Manuscript in preparation 62

65 Abstract Successful development of Plasmodium parasite in the mosquito is a prerequisite for the transmission of malaria. The parasite goes through a complex developmental cycle. Of the hundreds of ookinetes that are formed in the mosquito midgut, only a few are able to successfully transform into oocysts. This huge loss in numbers is attributed to the efficient killing of parasites by the mosquito immune system. However, once oocysts are formed, they appear to be resistant to the mosquito defenses. A single oocyst can produce thousands of sporozoites, the infective form of the parasite. During oocyst development, a thick capsule surrounds the parasite and appears to function as a protective cover. We hypothesized that interfering with capsule formation/assembly may lead to killing of the oocysts. In this study, we have identified a novel Plasmodium berghei oocyst capsule protein (PbCap380). We show by genetic analysis that this protein is essential for parasite development in the mosquito. In the absence of PbCap380, although oocyst formation is normal, the parasites do not survive in the mosquito and thus cannot be transmitted. We propose that targeting the oocyst capsule could provide new strategies to control malaria. 63

66 Introduction The malaria parasite life cycle is dependent on intricate association with its hosts. In the mosquito vector where sexual development takes places, the parasite undergoes a complex developmental program that involves crossing two distinct epithelial cell layers, midgut and salivary gland (Ghosh et al. 2000). Ookinetes that are formed within the blood meal traverse the midgut epithelium and transform into an oocyst. Interaction of the ookinete with basal lamina proteins is thought to trigger this transformation (Arrighi and Hurd, 2002). The oocyst remains sandwiched between the epithelial cells and the basal lamina for days. During this period the parasite grows in size and undergoes numerous nuclear division resulting in haploid nuclei (Howells and Davies 1971a,b, Canning and Sinden 1973). As development proceeds, a distinct capsule secreted by the oocyst separates the parasite from the mosquito tissue (Aikawa 1971). The molecular composition of the capsule remains unknown although oocyst capsule proteins have been speculated to interact with components derived from the mosquito (Adini and Warburg 1999). Recently a transglutaminase activity has been reported in oocysts and may function in crosslinking parasite and mosquito proteins (Adini et al. 2001). Soon after the parasite traverses the midgut, ookinete surface proteins cover the young oocyst. Thereafter, new surface proteins must be synthesized during oocyst development. As the ookinete traverses through the midgut, it elicits a potent mosquito immune response (Dong et al. 2006). Drastic reductions in parasite load occur within the 64

67 midgut lumen and during ookinete migration through the midgut epithelium (Gouagna et al. 1998). Known mechanisms of parasite killing are lysis of ookinetes during their invasion of the midgut epithelium (Vernick et al. 1995) and melanotic encapsulation of the parasites as they emerge on the basal side of the midgut (Collins et al. 1986). However, the oocysts appear to be more resistant to the mosquito defenses. The capsule probably plays an important role in protecting the developing parasite. At the end of parasite development, a single oocyst produces thousands of infective sporozoites, which are transmitted when the mosquito bites a vertebrate host. If one can eliminate those last few oocysts, we might be able to control malaria successfully. For instance, interfering with the formation of the capsule could lead to the oocysts being recognized and killed by the mosquito immune system. Here we report the functional characterization of a novel Plasmodium-specific capsule protein in P. berghei (PbCap380). PbCap380 is expressed only in the during oocyst development and localizes to the capsule surface. Targeted disruption reveals the function of the protein in protecting the oocyst, as parasites lacking PbCap380 do not survive to complete their development. As a result mosquitoes infected with PbCap380(-) parasites are completely devoid of sporozoites and are unable to transmit the parasite. Taken together, this is the first report of a protein specifically expressed on the oocyst capsule surface and is absolutely essential for parasite survival inside the mosquito vector. 65

68 Materials and Methods Parasite maintenance, mosquito infection were performed as described previously (see Chapter 2, Srinivasan et al., 2004) PbCap380 gene expression analysis. Total RNA was isolated with Trizol (Molecular Research Center) from mixed blood stage parasites, infected midguts (24hr, day 2, day 4, day 9, and day 14 after blood meal) and salivary gland sporozoites. PCR was performed using gene specific primers PbCap380F- 5 GAAATCACCATTTAATTTCTCCAATGGGT3 and PbCap380R- 5 TGTAGTTCGAAAAGGATGGTTTTGATTGT3. Plasmodium ribosomal RNA (rrna) (PbrRNA1-5 TGGGAGATTGGTTTTGACGTTTATGT3 and PbrRNA2-5 AAGCATTAAATAAAGCGAATACATCCTTAC3 ) was used for normalization (Arreaza G et al. 1991). Antibody synthesis and western blot analysis. A polyclonal anti-sera was raised against the highly charged repeat region of PbCap380 in rabbit. For detection of PbCap380 protein by western blot, we dissected 40 midguts from uninfected and day 15 infected mosquitoes. The protein lysates were run on a SDS-PAGE gel and detected using anti-pbcap380(1:1000) antibody and a horse raddish peroxidase (HRP) conjugated anti-rabbit secondary antibody (Sigma). Oocyst immunofluorescence. To detect PbCap380 in oocysts, infected midguts were 66

69 fixed in 4% paraformaldehyde for 2 h and were permeablized in PBS containing 0.2% Triton X100. The midguts were then incubated with the primary antibody (1:1000) and detected using a rhodamine conjugated anti-rabbit IgG. Pre-immune sera obtained from the same rabbit was used as a control. To count early oocysts, midguts were dissected and the leftover blood meal was carefully removed and the resulting midgut sheets were processed as mentioned above. Oocysts were counted by staining with anti-pbs21(1:1000) and anti-pbcsp (3D11, 1:1000) monoclonal antibodies and detected using FITC conjugated anti-mouse secondary antibody (Sigma). Confocal microscopy. Midguts were dissected 15 days after infection and processed as mentioned above. Oocysts were double labeled with anti-pbcap380 and anti- PbCSP antibodies and detected using rhodamine conjugated anti-rabbit and FITC conjugated anti-mouse secondary antibodies respectively. Immunoelectron microscopy. Infected midguts were fixed for 20 min at 4 C with 1% formaldehyde and 0.2% glutaraldehyde in 0.1 M phosphate buffer, ph 7.4. Fixed samples were washed, dehydrated, and embedded in LR White resin (Polysciences). After blocking the thin sections in phosphate-buffered saline containing 5% w/v nonfat dry milk and 0.01% v/v Tween 20 (PBTM), the grids were incubated with anti-pbcap380 primary antibody. Following labeling of the sections with 5-nm goldconjugated goat anti-rabbit IgG (Amersham Biosciences), samples were stained with uranyl acetate and lead citrate and then examined in a Zeiss CEM902 electron microscope (Oberkochen, Germany). 67

70 Generation of PbCap380 (-) parasites. For targeted disruption of the PbCap380 locus, a disruption plasmid was generated by amplification of a PCR fragment using primers IntgF (5'GCATGTAAAGTAATAAAACCATCTACA3') and IntgR (5'AGGTGTAAATAATGATATGAAACCT3') and P. berghei genomic DNA as template. Cloning into the P. berghei transfection vector (van Dijk et al. 1995) resulted in plasmid pcap380. The disruption plasmid was linearized at the unique NdeI site, transfected into P. berghei schizonts and disruptants were selected and cloned as described previously (Waters et al. 1997). To confirm disruption of the PbCap380 locus, integration-specific PCR was performed using the following primers, P1-5 TGTAGTTCGAAAAGGATGGTTTTGATTGTTT3,P2-5 ACATAC TTGATTTCAGCACCTTATCAGA3 and P3-5 GAGTTCATTTTACACAATCC3. Two independent PbCap380 (-) clones were obtained which were phenotypically identical. Results from one representative clone are discussed. Phenotypic analysis of PbCap380 (-) parasites. Development of wild type (WT) and PbCap380(-) parasites in mice was assessed by injecting a known number of parasites in Swiss Webster mice. The ability of parasites to differentiate into gametocytes and form male gametes (exflagellation) was assessed as described previously (Dearsly et al. 1990). Infected mice were fed to A.stephensi mosquitoes and the ability of the mutant parasites to form ookinetes (24 h) and oocysts (day 2- day 15) were examined microscopically. For counting ookinetes, blood fed midgut were dissected 24 h after infection and counted by Geimsa staining. Developing 68

71 oocysts were counted by staining the parasites as described above. Mature oocysts were counted on day 15 under a light microscope. Results Identification of Plasmodium berghei capsule protein 380 (PbCap380). The gene for PbCap380 was initially identified in a subtractive hybridization screen for differentially expressed transcripts in the oocysts (Srinivasan et al. 2004). The Cap380 gene can be identified exclusively in the Plasmodium genus and orthologues are found in all Plasmodium species for which gene sequences are available. All Plasmodium Cap380 genes are predicted to be transcribed from a single exon that codes for a putative signal sequence at the N-terminal (Fig.3.1A). The P. falciparum and P. vivax orthologues share 25% and 22% amino acid identity with the P. yoelli protein respectively (PlasmoDB accession numbers PfFC0905c, Pv095215, PY00597, PB and PB ). However several microdomains within the protein have up to 47% sequence identity between multiple Plasmodium species. These conserved regions could be important for protein function. The rodent Cap380 proteins have a highly charged repeat sequence that is absent in P. falciparum. Quantitative RT-PCR analysis shows that PbCap380 is specifically induced during oocyst development in the mosquito (Fig.3.1B). Gene expression mirrors oocyst development in the midgut where the oocyst attains its maximum size around day 12 after infection. More importantly, the gene is neither induced in blood stage parasites 69

72 nor in purified ookinetes and sporozoites (Fig.3.1B). This is consistent with the recent P. falciparum microarray examining gene expression during the parasite life cycle, which shows that Cap380 is not expressed in sporozoites and blood stage parasites (Le Roch et al. 2003). PbCap380 protein is expressed in oocysts. Plasmodium oocyst development in the mosquito is asynchronous. Western blot analysis and immunofluroscence labeling of day 15 infected midguts identified PbCap380 expression in oocysts (Fig.3.1 C and D). In order to analyze the precise time at which the protein is expressed, we labeled mosquito midguts with anti-pbcap380 antibodies at different times after infection. PbCap380 is expressed immediately after the ookinete transforms into an oocyst (36h) and continues throughout oocyst development till sporozoites begin to be released (day 15) (Fig.3.2A). However, the protein is expressed only in the oocysts and not in any other parasite forms (data not shown). PbCap380 localizes to the capsule of the oocyst. In early oocysts (day 3), we find both Pbs21 (ookinete surface protein) and PbCap380 localizing to the oocyst surface (data not shown). As development proceeds (day 5 and day6), Pbs21 appears to be internalized and PbCap380covers the oocyst surface (data not shown). Plasmodium berghei circumsporozoite protein (PbCSP) is a major sporozoite surface protein (Rogers et al. 1992). CSP also localizes to the inner surface of the oocyst wall and is required for oocyst differentiation (Menard et al. 1997). In order to determine the localization of PbCap380 relative to PbCSP, day 15 oocysts were double labeled with 70

73 anti-sera against PbCap380 and PbCSP. Though both the proteins are present in close proximity to each other, PbCap380 is present external to PbCSP (Fig.3.2B). Immunoelectron microscopy (IEM) confirmed the capsule localization of PbCap380 (Fig.3.4). PbCap380 labeling is also detected in regions where the mosquito tissues was originally in contact with the oocyst capsule but were detached as a result of the tissue sectioning (Fig.3.4). This is consistent with the notion of parasite capsule proteins interacting with mosquito-derived proteins. Fig.3.1. Identification of PbCap380 as a oocyst protein. A, PbCap380 is a single exon gene and codes for a 380kd protein. The rodent parasite proteins share 25% sequence identity with the human malaria parasites P. falciparum and P. vivax. However, there are micro-domains (shaded gray) within the protein that have up to 47% sequence identity 71

74 between multiple parasite species. The predicted signal peptide is indicated in black and the highly charged repeat sequence in the rodent parasite is hatched. B, Quantification of gene transcription. Quantitative RT-PCR was performed using cdnas prepared from blood stage parasites (Asx), in vitro cultured ookinetes (ook), midguts dissected at different times after infection (24 h 15 d) and sporozoites from salivary gland (SgSpz) of infected mosquitoes. Expression of PbCap380 is up-regulated during oocyst development in the mosquito. Peak mrna abundance occurs around 9 d after the mosquitoes ingest an infected blood meal. Three biologically independent quantitative RT-PCR experiments were performed and the pattern of gene expression was similar. Data shown is from one representative experiment. C, Western blot analysis. PbCap380 is present in day 15 infected midguts (Inf lane) but not in uninfected midguts (UInf lane). D, Indirect immunofluorescence of day 15 oocyst. P. berghei oocyst labeled with anti-pbcap380 antibodies detect the protein on the oocyst periphery. The staining pattern suggests a surface localization. The protein is not expressed in any other parasite forms (data not shown). Generation of PbCap380(-) parasites. PbCap380 is specifically expressed in the mosquito stages of parasite development. To further insights on the role of PbCap380, we disrupted the PbCap380 gene in parasites (Fig.3.4A). Independent clonal parasite lines generated from the transfection were confirmed for gene disruption by insertionspecific PCR (Fig.3.4B). Moreover, PbCap380(-) oocysts do not express PbCap380 (Fig. 3.4C). As expected the mutant parasites developed normally in the vertebrate host (data not shown). 72

75 Fig.3.2. PbCap380 protein can be detected throughout oocyst development. A, Indirect immunofluorescence of developing oocysts. PbCap380 expression begins soon after the formation of oocyst (36h) and continues all the way till the end of oocyst development (day 15). B, Confocal microscopy section of a day 15 oocyst. The parasite was double labeled with antibodies against PbCap380 (red) and PbCSP (green). A higher magnification of the inset from the merged image shows that PbCap380 is present on the surface of the oocyst, while CSP is below the surface 73

76 Fig.3.3. PbCap380 localizes to the oocyst capsule. A, Immunoelectron microscopy labeling with PbCap380 antibody. Immunogold labeling of oocyst capsule clearly localizes PbCap380 to the oocyst capsule. B, PbCap380 may interact with the mosquito tissue. Immunogold labeling is observed in regions of the mosquito tissue, which was originally in contact with the oocyst capsule. C, capsule, Oo, oocyst, Mq, mosquito 74

77 Fig.3.4. Targeted gene disruption of PbCap380. A, Schematic representation of the targeting strategy. The wild type PbCap380 genomic locus (WT) is targeted with an NdeIlinearized plasmid (pcap380) containing the 5 and 3 truncations of the PbCap380 open reading frame and the TgDHFR positive selection marker. Upon a single crossover event, the region of homology in duplicated, resulting in two truncated, non-expressed PbCap380 copies in the integrated locus (PbCap380(-)). The homologous regions in the disruption plasmid are shaded gray. Arrowheads indicate primer pairs used to confirm gene disruption. B, Integration-specific PCR analysis. Genomic DNA was prepared from drug resistant parasite clones and PCR was performed using the primer pairs indicated. The presence of the 2358bp integration-specific PCR product (P1/P3) but not the 2063bp WT locus-specific PCR product (P1/P2) in the PbCap380(-) lanes confirm gene disruption. Note that WT lanes show the presence of the wild type locus (P1/P2) as expected but not the integration locus (P1/P3). C, PbCap380 cannot be detected in the knockout parasite. Wild type and PbCap380(-) oocysts (day 3) were double labeled with Pbs21 and PbCap380. As expected 75

78 the knockout parasites do not express Cap380 but they express Pbs21. D, Morphology of day 15 wild type (WT) and PbCap380(-) oocysts. WT parasites develop normally and form mature, sporozoite-containing (arrow) oocysts, while PbCap380(-) oocysts die. Very rarely abnormal oocyst-like structures (arrow head) are observed in some PbCap380(-) infected mosquitoes. PbCap380 is required for oocyst development. To study the role of Cap380 in sexual stage development, the PbCap380(-) parasites were fed to mosquitoes. Formation of male and female gametocytes and the subsequent formation of male gametes were not affected in the knockout parasites (data not shown). The ability of the knockout parasites to form ookinetes in vivo was assessed in infected midguts dissected 24h after blood meal. PbCap380(-) parasites were able to transform into ookinetes in normal numbers (Table.3.1). However, when analyzed 15 days later, neither mature oocysts nor sporozoites could be detected (Table.3.1). In a few mosquitoes we observed what looked like small, disintegrating oocysts (Fig.3.4D). In contrast, WT parasite fed midguts contained numerous oocysts that produced infectious spororzoites (Table.3.1). As expected, mosquitoes carrying PbCap380(-) parasites were unable to transmit the parasite (data not shown). The absence of oocysts could be due to either of two possibilities, namely, PbCap380(-) ookinetes not being able to invade the midgut or the PbCap380(-) oocysts unable to survive in the mosquito. As the protein is not expressed in the ookinetes (data not shown), it is unlikely that it is involved in invasion. To test this, we analyzed the ability of the knockout ookinetes to traverse the midgut epithelium 76

79 and transform into oocysts. Mosquito midguts were dissected 3 days after infection and stained with an antibody against Pbs21. Our results show that PbCap380(-) ookinetes are able to invade the midgut epithelium and transform successfully into oocysts similar to wild type parasites (Fig.3.5). Table 3.1. PbCap380(-) parasites do not form mature oocysts. a, Ookinetes formed in vivo in the blood meal was counted from individual mosquitoes 24h after infection. Results are shown as average ookinetes/mosquito and was obtained from two wild type and three PbCap380(-) independent experiments. b. Oocysts were counted 15 days after infection. Results are shown as average oocysts/mosquitoand was obtained from four wild type and ten PbCap380(-) independent experiments. We did not find a single mature oocyst in PbCap380(-) infected mosquitoes. c. Sporozoites were prepared from midguts dissected 15 days after infection. Results are shown as average sporozoites/mosq. As expected mosquitoes infected with PbCap380(-) do not produce any sporozoites. PbCap380(-) oocysts do not survive in the mosquito. In order to understand the fate of PbCap380(-) oocysts, we examined midguts at different times after infection and counted the number of oocysts by staining with anti-pbs21 and anti-pbcsp antibodies. Pbs21 is expressed until 5 6 days after infection and PbCSP begins to be 77

80 Fig.3.5. PbCap380(-) oocysts do not survive in the mosquito. A, Immunofluorescence of developing oocysts. B, Quantitation. Midguts were dissected at different times after infectious blood meal and stained with a mixture of anti-pbs21 and anti-pbcsp antibodies. The number of labeled oocysts was counted for each time point. The highest number of parasites between all the time points is considered 100%. As development proceeds, the number of Cap380(-) oocysts reduces drastically, presumably killed by the mosquito immune system, whereas the number of Cap380(+) oocysts do not change substantially. Results are pooled from two independent experiments. For each time point 5-15 midguts were counted. expressed from about day 6 (Srinivasan and Jacobs-Lorena, unpublished). While both PbCap380(+) and PbCap380(-) parasites transform into oocysts efficiently, the 78

81 number of Cap380(-) oocysts declines as development proceeds (Fig.3.5). It has been speculated that the ookinete surface proteins Pbs21 and Pbs25 in addition to their role in mosquito-parasite interactions (Siden-Kiamos et al. 2000, Vlachou et al. 2001), may also play an important role in protecting the ookinete (Saxena et al. 2006). While the early oocyst is still covered by ookinete surface proteins, their synthesis stops there after. As the oocyst matures, new capsule proteins replace existing surface proteins. PbCap380 is an essential component of the oocyst capsule and its absence probably results in the developing oocysts being killed by the mosquito. Discussion A drastic reduction in parasite number occurs as it transits from gametes to ookinetes to oocysts. Hundreds of ookinetes are formed in the midgut lumen, but only a few are able to develop into oocysts successfully. This is most likely due to a strong immune response elicited by the mosquito that recognizes the invading parasite inside the epithelium (Vernick et al. 1995, Dong et al. 2006) as well as when it emerges on the basal side of the midgut (Collins et al. 1989). Both pro- and anti-parasite molecules are known to determine the success of infection (Osta et al. 2004, Vlachou et al. 2005). These studies show that pro-parasite molecules such as a C-type lectin indeed protect the parasite (Osta et al. 2004) possibly by covering the parasite and making it inaccessible to the anti-parasite molecules. All such effector molecules identified to date have been shown to act on the ookinetes. The parasites that succeed in transforming into oocysts seem to somehow evade the immune system and are able to 79

82 form the infective sporozoites. What is not clear is how the oocysts are able to evade the immune molecules or why they fail to induce them. Ookinete surface proteins interact with mosquito basal lamina proteins (Vlachou et al. 2001, Dessens et al. 2003, Mahairaki et al. 2005) and this interaction is thought to trigger the transformation of the ookinete into an oocyst. The new oocyst is surrounded by ookinete surface proteins that are probably replaced by oocyst-specific proteins. The capsule, which is secreted soon after oocyst formation, surrounds the entire parasite and separates it from the mosquito tissues. This might also function in protecting the developing parasite from the surrounding harsh environment. The parasite has to acquire nutrients for its developing needs. The capsule is a porous structure capable of letting amino acids and nucleotides to pass through and enter the parasite (Vanderberg et al. 1967). The molecular nature of the capsule remains largely unknown. Here, we identify PbCap380 as the first member of the oocyst capsule. PbCap380 is specifically expressed during oocyst development and localizes to the oocyst capsule. Genetic studies identify an essential role for PbCap380 in protecting the developing parasite. The intricate association of the capsule with mosquito tissue raises an interesting possibility of interactions between capsule and mosquito proteins. This notion is supported by the observation of transglutaminase activity being detected only in infected midguts (Adini et al 2001) and the ability of basal lamina proteins to bind to the oocyst capsule (Adini and Warbug 1999). If parasite caspule and mosquito 80

83 proteins do interact, detachment of the oocyst from the mosquito midgut would result in some capsule components associated with the mosquito tissue and vice versa. PbCap380 is also seen attached to the adjacent mosquito tissue at junctions, which was originally in contact with the oocyst, suggesting that capsule proteins may interact with mosquito-derived components. As discussed earlier, the capsule might function to protect the parasite. We speculate that capsule proteins such as PbCap380 interact with one another as well as with mosquito proteins. Such interactions with mosquito proteins may coat the oocyst surface and render the parasite invisible to the mosquito immune system. If this were true, interfering with this interaction would result in efficient killing of the parasite. In support of this hypothesis, oocysts lacking a functional PbCap380 protein do not survive in the mosquito. Disruption of PbCap380 may interfere with capsule assembly resulting in a naked oocyst, which is easily recognized and killed by the mosquito. Hence mosquitoes infected with Cap380(-) parasites do not produce sporozoites and can not transmit malaria. Elucidating the mechanism by which the PbCap380(-) oocysts are killed will be the focus of future investigation. We have identified PbCap380 as an important component of the oocyst capsule. Our data suggests a role for the capsule in protecting the parasite. Parasites lacking PbCap380 are probably killed by the immune system of the mosquito. As a result mosquitoes infected with PbCap380(-) parasites are unable to transmit malaria. Our study shows that interfering with a single component of the capsule is very effective 81

84 in killing the parasite and identifies Plasmodium oocyst as a potential target to interfere with disease transmission. 82

85 CHAPTER 4 Distinct Roles for Rhomboid1 and Rhomboid3 proteins in Malaria Pathogenesis and Plasmodium Development Prakash Srinivasan, Hisashi Fujioka and Marcelo Jacobs-Lorena Reference: Manuscript in preparation 83

86 Abstract Malaria parasite invasion of host-cells involves recognition and interaction with multiple cell-surface receptors. This process utilizes a wide variety of parasite surface proteins, most of which are specific to the particular invasive form of the parasite, while some are shared among multiple forms. For a parasite to enter the cell, it has to somehow dissociate itself from the receptor. One mechanism by which it does so is by shedding its surface ligands using specific enzymes. Rhomboid proteins are a new family of serine proteases that cleave certain cell-surface proteins within their transmembrane domain. Here we identify two Plasmodium berghei rhomboid proteases that play distinct roles in the three invasive stages of the parasite. Disruption of genes encoding rhomboid1 and rhomboid3 proteins interferes with the parasite s ability to invade the mosquito midgut. Rhomboid3 is also essential for sporozoite formation. Rhomboid1 on the other hand appears to play an important role in all the three invasive stages of the parasite. Most of the animals infected with rhomboid1 disruptants are able to clear the parasites efficiently and are protected from a subsequent lethal P. berghei wild type challenge. Our studies suggest a role for rhomboid proteins in all invasive stages, identifying them as attractive targets for disease intervention. 84

87 Introduction For successful development and transmission, malaria parasite Plasmodium has to invade multiple cell types both in the mammalian host and in the mosquito vector. Much of our knowledge about the molecular mechanisms of invasion comes from the study of P. falciparum merozoite invasion of red blood cells (RBCs). RBC invasion involves an initial attachment followed by re-orientation and entry of the parasite into the host cell (Dvorak et al. 1975). There are two main classes of parasite surface molecules, the GPI-anchored proteins such as the merozoite surface protein family (MSP) (Siddiqui et al. 1987) and transmembrane domain-containing proteins such as AMA1 (Klotz et al. 1989, Mitchell et al. 2004), erythrocyte binding-like family (EBL) (Adams et al. 1992, Adams et al. 2001) and reticulocyte binding-like family proteins (RBL) (Galinski et al. 1992, Rayner et al. 2000). A few of the host-cell receptors to which these ligands bind have been identified (Miller et al. 1975, Dolan et al. 1994, Maier et al. 2003, Sim et al. 1994). In case of ookinetes and sporozoites, the other two invasive forms, motility is an important feature of the parasite. The ookinetes that form within the mosquito blood meal actively migrate towards the midgut epithelium. They invade multiple epithelial cells before exiting on the basal side, facing the hemocoel, and transforming into an oocyst (Han et al. 2000). Similarly, the sporozoites that are released into the mosquito hemocoel have to first invade the salivary gland and subsequently, after being delivered in the vertebrate host, travel through the blood stream before infecting 85

88 vertebrate hepatocytes. During this process called gliding motility, the parasite uses surface proteins to attach to receptors. Even though merozoites do not have a typical gliding motility, all the three invasive forms utilize the same actin-based motility for entry into the host cell. The thrombospondin-related anonymous protein (TRAP) family homologs are found in all the three invasive forms and are required for motility and the host cell invasion (Sultan et al. 1997, Dessens et al. 1999, Baum et al. 2006). The extracellular domains of TRAP interact with host-cell receptor, while the cytoplasmic tail links to the actin-myosin cytoskeleton (Kappe et al. 2004). As the parasite glides, the surface ligands that are engaged with the receptors translocate towards the posterior end. Dissociation of these interactions by proteolytic processing is thought to be important, as this enables the parasite to move forward (Carruthers et al. 2000, Howell et al. 2003, Coppi et al. 2005). MIC2, the TRAP homologue in Toxoplasma, another Apicomplexan parasite, is cleaved within its membrane domain releasing the receptor-binding domain from the surface (Carruthers et al. 2000). Plasmodium merozoite TRAP (MTRP) and TRAP are also cleaved in a similar manner (Baum et al. 2006, Srinivasan et al. unpublished). Inhibitors that block this processing prevent host cell infection (Srinivasan et al. unpublished). Rhomboid-family (ROM) proteins are a new class of serine proteases that cleave their substrates within their membrane domain (Urban et al. 2001, Freeman 2004). Processing requires the presence of helix-destabilizing residues in the membrane domain of substrates (Urban and Freeman 2003). The genomes of Plasmodium and Toxoplasma contain multiple rhomboid-family proteins (Dowse and Soldati 2005). 86

89 Further, the Apicomplexan surface proteins including the EBL and RBL proteins, AMA1, TRAP and their homologues contain helix-destabilizing residues in their membrane domain (Dowse and Soldati 2005). Indeed, Toxoplasma ROM5 localizes to the posterior end of the parasite and cleaves MIC2 within the transmembrane domain in in vitro assays (Brossier et al. 2005, Dowse et al. 2005). Plasmodium does not have a ROM5 homologue but ROM4, is able to cleave EBA175 (O Donnell et al. 2006), an EBL family protein involved in binding to erythrocytes (Sim et al. 1994). Processing of EBA175 is essential for parasite invasion (O Donnell et al. 2006). These studies however do not show the role of rhomboid proteins in the parasite. Here we demonstrate that two Plasmodium berghei rhomboid proteins ROM1 and ROM3 play distinct roles at multiple stages of parasite invasion. Unlike ROM1, ROM3 is essential for sporozoite formation. Both ROM1 and ROM3 are involved in ookinete invasion of the mosquito midgut, while ROM1 is also required for efficient hepatocyte infection. Even though ROM1 deficient parasites are able to infect the vertebrate host, their growth is impaired, possibly as a result of poor invasion of merozoites into RBCs. As a result mice are able to clear the infection efficiently. Further, animals immunized with ROM1 deficient parasites are protected from a subsequent lethal wild type parasite challenge. Our findings reveal an important function of rhomboid family proteins in the malaria parasite life cycle. Therefore rhomboid proteases offer a unique target for interfering with both disease causing and disease transmitting forms of the parasite. 87

90 Materials and Methods Parasite maintenance and mosquito infections were performed as described previously (Chapter 2, Srinivasan et al., 2004) ROM1 antibody synthesis and Immunofluorescence assays. A rabbit polyclonal anti-sera was raised against the N-terminal 52 amino acids of PbROM1. P.berghei schizonts, merozoites and sporozoites (isolated from mosquito midguts and salivary glands) were fixed in methanol and labeled with anti-rom1 (1:500) anti sera. Schizonts and merozoites were also labeled with mouse anti-ama1 antibody, while midgut and salivary gland sporozoites with mouse anti-csp antibody. Slides were then incubated for 1 h with Alexa Fluor 488 anti-rabbit IgG and Rhodamine antimouse IgG secondary antibodies. After washing, images were visualized in a Leica upright fluorescent microscope at 100X objective and images were captured using the SPOT camera. Generation of PbROM1 and PbROM3 disruptants. For targeted disruption of the PbROM1 and PbROM3 locus, a disruption plasmids were constructed by amplification of PCR fragments using primers, PbROM1F(5'CCATACATTAGCAGAGTATAGGGA3'),PbROM1R(5'ACTTGCAC CCACTTTTATTGTAC3'),PbROM3F(5 CGGACCATCTGTCAAGTACCTGCCT3 ) and PbROM3R(5 ATCAACAGTTGGAGAAAAACTTAC3 ) respectively using P. berghei genomic DNA as template. Cloning into the P. berghei transfection vector 88

91 (van Dijk et al. 1995) resulted in plasmid pcap380. The ROM1 disruption plasmid was linearized at the unique NdeI site, and ROM3 plasmid at the unique XmI site and transfected into P. berghei schizonts. Disruptants were selected and cloned as described previously (Waters et al. 1997). To confirm disruption of the PbROM1 and PbROM3 locus, integration-specific PCR was performed using specific primer combinations, P1(5 CGAGCAACAATGTCTGAC3 ),P2(5 GAGTTCATTTTACACAATCC3 ), P3(5 TAATACGACTCACTATAGGGAGA3 ),P4(5 TGTTACCTGGGACTACTG AGCTGA3 ),P5(5 AGCACAAATCGAGGCAAATCCATA3 ), and P6(5 TAACAATGATATCATCACA3 ) Phenotypic analysis of PbROM1 and PbROM3 disruptants. The ability of the parasites to differentiate into gametocytes and form male gametes (exflagellation) was assessed as described previously (Dearsly et al. 1990). Infected mice were fed to A.stephensi mosquitoes and the ability of the disruptant parasites to form ookinetes (24 h) and oocysts (day 2- day 15) were examined microscopically. For counting ookinetes, blood fed midgut were dissected 24 h after infection and counted by Geimsa staining. Mature oocysts were counted on day 15 under a light microscope. Midgut and salivary gland sporozoites were counted on day 25. Sporozoite infection of the vertebrate host. Sporozoites isolated from salivary glands were counted using a hemocytometer. Animals were injected intra venously with 150, 1000 or sporozoites. For testing efficiency of liver infection, animals 89

92 were scarificed h after sporozoite injection. Total RNA was prepared from wild type and ROM1(-) infected mice using Trizol reagent. P. berghei 18S rrna was quantified using primers (PbrRNA1-5 TGGGAGATTGGTTTTGACGTTTATGT3 and PbrRNA2-5 AAGCATTAAATAAAGCGAATACATCCTTAC3 ) as described before (Arreaza G et al. 1991) and was normalized using mouse GAPDH. Infection efficiency was also assayed by monitoring the pre-patent period of blood stage infection after sporozoite injection. Blood stage infection. For assaying blood stage infections, animals were injected with either wild type or disruptant sporozoites (intra venous) or infected red blood cells (irbcs) (intra venous or intra peritoneous). Number of infected RBCs was counted on Geimsa stained blood smears everyday till end of the experiment. Parasite challenge. Mice were infected with ROM1(-) parasites as described above. Parasitemia was checked everyday until at least 30 days after the last ROM1(-) parasites was detected. To confirm complete parasite clearance, 3x10 7 RBCs from these animals were injected into naïve mice. ROM1(-) immunized animals were then challenged with 10 6 wild type irbcs intra venously. Parasitemia was followed as described above. Results ROM1 and ROM3 are conserved in all Plasmodium species. Plasmodium berghei ROM1 (PbROM1) was initially identified by a subtractive hybridization screen for 90

93 genes expressed during parasite development in the mosquito. ROM3 was identified based on homology searches and is highly induced in P. falciparum gametocytes (Le Roch et al. 2003). Both genes encode multi-membrane proteins and have a conserved Asparagine, Glycine-X-Serine and Histidine rhomboid motif in the N-terminal onethird region of their membrane domains (Fig.4.1A). Both PbROM1 and PbROM3 homologues are highly conserved between rodent and human malaria species; however, sequence identity between the two rhomboids within a species is very limited (Fig.4.1B). This indicates an independent evolution of the rhomboid genes and suggests distinct functions for these two proteins in parasite development. Fig.4.1. Identification of P. berghei rhomboid 1 and rhomboid 3. A, Conservation of catalytic residues between Plasmodium ROM1 and ROM3 and Drosophila ROM1. The Asparagine-Serine-Histidine catalytic triad (indicated by *), which is essential for serine protease activity in Drosophila, is conserved in Plasmodium ROM1 and ROM3. The other surrounding amino acids shown are important for rhomboid protein function in Drosophila. All 91

94 the catalytic residues are predicted to be present within the transmembrane domains (shaded gray). B, Sequence identity between Plasmodium rhomboids. While there is very high sequence conservation among homologues for each rhomboid, there is minimal conservation between different rhomboid proteins of the same Plasmodium species. Pb, Plasmodium berghei, Py, Plasmodium yoelli, Pf, Plasmodium falciparum, Dm, Drosophila melanogaster ROM1 protein is expressed in merozoites and salivary gland sporozoites. We performed indirect immunofluorescence to determine PbROM1 protein expression in blood-stage and mosquito-stage parasites. ROM1 protein has a punctate distribution in segmented schizonts (mature) and localizes to the apical end of released merozoites (Fig.4.2A). Further ROM1 co-localizes with apical membrane antigen 1 (AMA1), a micronemal protein (Fig.4.2A). A similar punctate pattern of protein expression and accumulation at the apical end is observed in salivary gland invaded sporozoites (Fig.4.2B), suggesting a micronemal localization in this parasite form as well. ROM1 labeling is also associated with the cell surface as seen by co-localization with circumsporozoite protein (CSP) (Fig.4.2B). ROM1 transcript was isolated from the mature oocysts (containing sporozoites). However, we did not observe fluorescence in midgut sporozoites, suggesting that ROM1 protein expression may be posttranscriptionally regulated. ROM1 expression in the only other invasive form, the ookinete, has yet to be determined. ROM1 expression in the critical invasive forms combined with the micronemal localization indicates that it may function during invasion. 92

95 Fig.4.2. ROM1 is expressed in merozoites and sporozoites. PbROM1 protein expression was assayed by indirect immunofluorescence. A, IFA of fully segmented schizonts and released free merozoites, double labeled with anti-rom1 (green) and anti-ama1 (red) antibodies. ROM1 is found associated with the micronemal protein AMA1. DAPI is not shown for clarity. B, IFA of midgut and salivary gland invaded sporozoites double labeled with anti- ROM1 and anti-csp antibodies. ROM1 expression in salivary gland sporozoites is consistent with micronemal localization. ROM1 is present on the surface as seen by co-localization with CSP. DAPI is shown in blue. ROM1, rhomboid 1, AMA1, apical membrane antigen 1, CSP, circumsporozoite protein ROM1 and ROM3 gene disruption. To study the role of ROM1 and ROM3 in Plasmodium development, we undertook genetic studies by site-directed disruption of protein function (Fig.4.3A). A drug marker was inserted in the open reading frame (ORF), resulting in the disruption of the protein by truncation. Independent parasite clones were selected from the successful transfections. Phenotypes of each disruptant 93

96 clone were identical and the results from one representative clone each is presented. Gene disruption was confirmed by insertion-specific PCR that identifies the disrupted locus from the wild type locus (Fig.4.3B). ROM1 plays an important role in ookinete invasion of the midgut epithelium. We examined the function of ROM1 in ookinetes by feeding ROM1(-) parasites to mosquitoes. The efficiency of ookinetes to invade the midgut was assessed by counting the number of oocysts formed. In several experiments we consistently found a strong reduction in oocyst numbers in spite of the development of similar number of ookinetes in both wild type (WT) and ROM1(-) mosquitoes (Table.4.1). Both WT and ROM1(-) oocysts develop normally and produce similar number of sporozoites (Table.4.1). This indicates that ROM1(-) ookinetes are impaired in their capacity to invade the midgut epithelium. The role of ROM1 in salivary gland invasion however is not clear. We did not observe a consistent difference in invasion efficiency between WT and ROM1(-) sporozoites in our assays (data not shown). Table 4.1. ROM1 and ROM3 proteins are required for midgut invasion. Number of ookinetes in the blood meal was counted in Geimsa stained smears od midgut contents 24 h after infection. For ookinete counts, 5 10 midguts were analyzed for each experiment. Oocysts were counted on day 15. Inhibiton of oocyst formation was calculated 94

97 using median oocyst numbers. For oocyst counts, mosquitoes were analyzed for each experiment. Mean sporozoites per oocyst were calculated using the mean number of oocysts. Sporozoites from mosquitoes were counted in each experiment. Numbers in parenthesis indicate the number of independent experiments that were performed to obtain the pooled data. Fig.4.3. ROM1 and ROM3 gene disruption. A, Schematic representation of the targeting strategy. The wild type ROM1 genomic locus (WT) is targeted with an NdeI-linearized plasmid (prom1) containing the 5 and 3 truncations of the ROM1 open reading frame and the TgDHFR positive selection marker. Upon a single crossover event, the region of homology in duplicated, resulting in two truncated, non-expressed ROM1 copies in the integrated locus (PbROM1(-)). The homologous regions in the disruption plasmid are shaded gray. Arrowheads indicate primer pairs used to confirm gene disruption. A similar strategy was used to disrupt the ROM3 gene with a XcmI-linearized plasmid. B, Integration-specific PCR analysis. Genomic DNA was prepared from drug resistant parasite clones and PCR was performed using the primer pairs indicated. For ROM1 disruption, the presence of the 1.2 kb integration-specific PCR product (P3/P2) but not the 1.7 kb WT locus-specific PCR product (P1/P2) in the PbROM1(-) lanes confirm gene disruption. Note that WT lanes show the 95

98 presence of the wild type locus (P1/P2) as expected but not the integration locus (P3/P2). For ROM3, primers P4/P6 (942 bp) and P4/P5 (1192 bp) were used to detect the correct integration and WT locus respectively. ROM1(-) sporozoites are defective in hepatocyte infection. Next we tested the role of ROM1 in liver infection. For this we injected WTand ROM1(-) salivary gland sporozoites intravenously in mice. The efficiency of infection was dose dependent and mice infected with ROM1(-) parasites showed a consistent delay in the pre-patent period by 1 d or more compared to wild type infected mice (Fig.4.4A). Infection was also determined by quantifying P. berghei 18S rrna levels in the infected liver 36-40h after sporozoite injection. We observed a 63% decrease in parasite load in mice infected with ROM1(-) parasites (Fig.4.4B). This suggests that ROM1 is required for efficient hepatocyte infection. ROM1 deficient parasites are impaired in blood stage infection. As ROM1 is required for efficient infection in at least two of the three invasive forms namely ookinete and sporozoite, we speculated that it might function in blood-stage infection as well. To test this hypothesis, we followed the growth of the parasite in mice infected with either sporozoites or blood-stage parasites. We found that parasitemia develops slower in animals infected with ROM1(-) parasites compared to WT infected mice (Fig.4.5). This phenotype is observed in animals infected by injection of sporozoites as well as bypassing liver invasion by injecting infected RBCs (irbcs) (Fig.4.5). As ROM1 appears to function in all the 96

99 three invasive stages of the parasite, we speculate that it might be involved in invasion of the parasite into the respective host cells. Fig.4.4. ROM1(-) salivary gland sporozoites are defective in mouse hepatocyte infection. A, Prepatent period of blood infection is longer in mice injected with ROM1(-) sporozoites. a. Mice were injected intra-venously with the indicated number of wild type (WT) or ROM1(-) salivary gland sporozoites. b. Number of mice infected/ number of mice injected with sporozoites. c. Number of days between sporozoite injection and the appearance of blood stage parasites upon examination of at least 25,000 RBCs. B, ROM1(-) sporozoites have decreased infectivity of the liver. Mice were injected with either 10 3 P. berghei wild type or ROM1(-) sporozoites (spz). Infection efficiency was assayed by measuring parasite load in the liver after 36-40h. Parasite load was determined by qrt-pcr. Results are expressed as mean± s.d. of rrna copy number from 4 mice per group. 97

100 Fig.4.5. ROM1(-) blood-stage parasites grow at a slower rate. Mice were injected with P. berghei wild type (WT) or ROM1(-) salivary gland sporozoites intravenously or with infected RBCs (irbcs) intraperitonealy. Parasite growth was monitored on Geimsa stained blood smears. The number of parasites used for infection is indicated inside each graph. For each experiment we used 3-4 mice. spz, sporozoite, irbc, infected red blood cell ROM1 infected mice survive longer. Mice infected with ROM1(-) parasites survived much longer than those infected with WT parasites (Fig.4.6). Animals infected with ROM1(-) parasites do reach peak parasitemia of greater than 35% similar to WT parasites. At such high parasitemia 100% of the animals infected with WT parasites succumb to anemia. But at least 80% of animals infected with ROM1(-) parasites survive and the majority of them eventually clear the parasites from their blood stream. The longer infections in mice presumably due to the reduced invasion of merozoites into RBCs seem to trigger an effective immune response against the parasite. 98

101 Fig.4.6. ROM1(-) infected animals survive longer. Swiss Webster mice were injected with the indicated number of salivary gland sporozoites. Mice survival was followed for one month. The numbers in parenthesis indicates the number of animals used in each experiment. Survival of animals injected with ROM1(-) irbcs is similar (data not shown). Immunization with ROM1(-) parasites protect mice form lethal WT P. berghei challenge. Next we wanted to test if the immune response generated against ROM1(- ) parasites could protect these animals from a subsequent wild type infection. Mice immunized with ROM1(-) parasites were challenged intravenously with 10 5 WT irbcs at least 30 d after the last parasite was detected. Parasitemia was determined by counting the number of irbcs in Geimsa stained blood smears (Fig.4.7). In this experiment very low level of parasitemia is detected at some point in 5/7 mice but all of them eventually clear the parasites and become infection-free (Fig.4.7). We speculate that these animals develop immunity against multiple parasite proteins and that contributes to the effective clearance of wild type parasites. 99

102 Fig.4.7. ROM1(-) immunization protects against WT parasite challenge. Animals were immunized with ROM1(-) parasites and were challenged intravenously with 10 5 WT P. berghei irbcs at least 30 days after the last ROM1(-) parasite was observed. The ability of immunized animals to clear the WT parasites was measured with Geimsa stained blood counting the first day after challenge as day 1. To assess protection, 3*10 7 RBCs form all the parasite-free animals were transferred into naïve mice. All the animals that were completely protected were re-challenged (indicated by arrow) again on day 33 and the parasitemia was followed. 100

103 Fig.4.8. ROM3 is required for sporozoite formation. Midguts of mosquitoes infected with WT, ROM3(-) and ROM3 revertant (ROM3rev) parasites were examined microscopically analyzed on day 15 after an infectious blood meal. WT and ROM3rev oocysts show the typical sporoblast formation and sporozoite budding, while ROM3(-) do not. Accumulation of vacuoles in ROM3(-) oocysts are readily observed. ROM3(-) oocsyst do however express CSP (ROM3(-)-CS), suggesting that the defect in sporozoite formation is independent of CSP expression. Sb, sporoblast, S, sporozoite, Vac, vacuoles ROM3 is required for ookinete invasion of the midgut. To study the role of ROM3 in parasite development, ROM3(-) parasites were fed to mosquitoes. The number of oocysts formed is a direct measure of invasion efficiency. In several independent experiments we find that ROM3(-) ookinetes were severely compromised in their ability to form oocysts (Table.4.1). This suggests that ROM3 function is required for normal invasion. Further, in several experiments we found a low prevalence of 101