I Malaria parasite metabolite and mosquito vector dynamics Erica Erlank A dissertation submitted to the Faculty of Science, University of the Witwatersrand, Johannesburg, in fulfilment of the requirements for the degree of Master of Science in Medicine Johannesburg, 2021 II Declaration I declare that this Dissertation is my own, unaided work. It is being submitted for the Degree of Master of Science (MSc Med) at the University of the Witwatersrand, Johannesburg. It has not been submitted before for any degree or examination at any other University. ________________________ (Signature of candidate) 15th day of September 2021 EERLANK III Publications arising from this research project *Due to the skills I have obtained during my MSc project. I was involved in other publications that arised from infection studies I participated in. Appendix A Reader J, van der Watt M, Taylor D, Le Manach C, Mittal N, Ottilie S, Theron A, Moyo P, Erlank E, Nardini L, Venter N, Lauterbach S, Bezuidenhout B, Horatscheck A, van Heerden A, Spillman NJ, Cowell AN, Connacher J, Opperman D, Orchard LM, Llinas M, Istvan ES, Goldberg DE, Boyle GA, Calvo D, Mancama D, Coetzer TL, Winzeler EA, Duffy J, Koekemoer LL, Basarab G, Chibale K, Birkholtz LM. 2021. Multistage and transmission-blocking targeted antimalarials discovered from the open-source MMV Pandemic Response Box. Nature Communications. 12:269. Role in the publication: I was responsible for coordinating infection feedings as well as conducting the standard membrane feeding assays, maintaining infected mosquitoes, midgut dissections of mosquitoes, data collection and analysis. Presentations arising from this research project Presentation was accepted for the H3D Symposium that was due to take place in March 2021 but it has been postponed till 2022 due to COVID-19 travel restrictions. IV Abstract Plasmodium falciparum makes its human host more attractive for its vector companion, the Anopheles mosquito. Once in the host, the malaria parasite synthesizes isoprenoids via the 2-C-methyl-D-erythritol 4-phosphate pathway. The precursor for isoprenoids in this pathway is known as (E)-4-hydroxy-3-methyl-but-2- enyl pyrophosphate (HMBPP). HMBPP activates red blood cells to release volatile organic compounds that acts as an attractant to the mosquito to stimulate blood feeding. Only a few anopheline mosquito species are known to transmit the human malaria parasite. These mosquito can be classified as either major or minor vectors, depending on their impact on malaria transmission. The main African malaria vectors are responsible for 95% of malaria cases while minor vectors contribute to the remaining 5%. During this study the aim was to evaluate if HMBPP influences P. falciparum susceptibility, feeding rate and attraction in major, minor and non-vector mosquito species. The standard membrane feeding assay was used to artificially infect mosquitoes and evaluate feeding rate, whereas attraction assays was conducted with a dual choice chamber. Results revealed that both the major vector, An. gambiae s.s., and the minor vector, An. merus, had an increase in both prevalence and intensity of oocysts. An increase in feeding rate and attraction was also observed for some vectors, in the presence of HMBPP. These findings could play an important role in understanding the role of parasite volatiles in malaria transmission. V Acknowledgements I am indebted to Prof. Lizette L. Koekemoer and Dr. Noushin Emami whose continued guidance enabled me to successfully complete this study. The patience and expertise with which they have guided me is truly appreciated. I would like to thank the following funders of this study: National Research Foundation/DST South African Research Chairs Initiative and the Strategic Health Innovation Partnerships (SHIP) of the South African Medical Research Council with funds received from the South African Department of Science and Technology. The shared data and assistance with field work undertaken in this project would not have been possible without the dedication by the SIT team and Dr. Givemore Munhenga. The generosity and assistance of the team at the University of Pretoria, who supplied the infected blood for many of the assays conducted in this study, is greatly appreciated and a special thanks to Prof. Lyn-Marie Birkholtz and Dr. Janette Reader for their continued support. To my creator, my family and husband, thank you for your encouragement and continued support. VI TABLE OF CONTENTS TITLE PAGE i DECLARATION ii PUBLICATIONS AND POSTERS/ CONFERENCES ARISING FROM THIS STUDY iii ABSTRACT v ACKNOWLEDGEMENTS vii TABLE OF CONTENTS viii LIST OF FIGURES ix LIST OF TABLES xii NOMENCLATURE/ABBREVIATIONS xiv CHAPTER ONE – INTRODUCTION 1.1 Malaria transmission 1 1.2 Objectives and hypotheses 4 1.3 Literature review 5 1.3.1 Mosquito vector dynamics and seasonal variation of species 5 1.3.2 Mosquito blood feeding and attraction behaviour 8 1.3.3 Parasite host manipulation 10 1.3.4 Parasite susceptibility of different mosquito vectors 14 VII CHAPTER TWO – MATERIALS AND METHODS 2.1. Sampling of wild mosquitoes 17 2.2. Wild caught mosquito identification 18 2.2.1. PCR species identification and seasonal trend 18 2.2.2. Data analysis 20 2.3. Wild caught mosquito infection profiling with the Enzyme linked immunosorbent assay (ELISA) 20 2.4. Attraction and feeding rate 23 2.4.1. Biological material 23 2.4.2. Mosquito attraction and dual choice chamber assays 24 i. Evaluation of choice chamber 24 ii. Attraction behaviour of major-, minor- and non-vectors in the presence of HMBPP 25 iii. Data analysis 26 2.4.3. Mosquito feeding rate and HMBPP 27 2.4.4. HMBPP Dose response evaluation 28 2.5 Artificial infection of different mosquito vector species with Plasmodium falciparum gametocyte culture 29 2.5.1. Infection through SMFA 29 2.5.2. Midgut dissections 31 2.5.3. Oocyst prevalence and intensity data capture and analysis 32 CHAPTER THREE – Results 3.1 Mosquito vector dynamics 3.1.1 Species identification and seasonal variation 33 3.1.2 Wild caught mosquito infection profiling 38 3.2 Mosquito blood feeding behaviour, attraction and parasite host manipulation 3.2.1 Mosquito attraction and HMBPP 39 3.2.1.1 Choice chamber evaluation 39 3.2.1.2 HMBPP and attraction of different mosquito vectors40 3.2.2 Feeding rate 44 VIII 3.2.3 HMBPP Dose response 45 3.3 Susceptibility of mosquito vector species to the P. falciparum parasite 3.3.1 Plasmodium falciparum susceptibility of major vectors 47 3.3.2 Plasmodium falciparum susceptibility of a minor vector 51 3.3.3 Plasmodium falciparum susceptibility of a non-vector 52 CHAPTER FOUR: General Discussion 4.1 Seasonal variation and vector dynamics 4.1.1 Species identification and seasonal variation 54 4.1.2 Wild caught mosquito infection profiling 55 4.2 Parasite metabolite (HMBPP) and mosquito behaviour 4.2.1 Mosquito attraction and HMBPP 56 4.2.2 Mosquito feeding behaviour in the presence of the parasite metabolite 57 4.2.3 Dose response evaluation 58 4.3 Mosquito susceptibility 61 CHAPTER FIVE: General Conclusion 58 CHAPTER SIX: References 62 CHAPTER SEVEN: Appendices 77 IX LIST OF FIGURES Chapter One Figure 1.1: Head and mouthparts of a female anopheline mosquito……………….10 Figure 1.2: Isoprenoid biosynthesis through the mevalonate and MEP pathways. The mevalonate pathway is found in higher eukaryotes whereas the MEP pathway is used by most eubacteria and apicomplexan protozoa……..13 Chapter Two Figure 2.1.a & b: Resting adult mosquitoes being collected from a clay pot and the search for resting adult mosquitoes in tyres at a sentinel site in Mamfene, KwaZulu Natal….………………….18 Figure 2.2: Egg laying vial containing a female Anopheles mosquito……………….23 Figure 2.3.a: Dual choice chamber with dimensions of the main chamber and two arms ……………………………………………………………………25 Figure 2.3.b: Dual choice chamber evaluation with sugar water and cow blood…...24 Figure 2.3.c: Mosquitoes entering the preferred arm of the choice chamber……….24 Figure 2.4: Set up for the standard membrane feeding assay with feeding cups placed under the feeding system with circulating water (37̊C) connected to the feeders…………………………….……………………..28 Figure 2.5.a: Microscopic image of an unfed female anopheline mosquito…………28 Figure 2.5.b: Microscopic image of a fully fed female anopheline mosquito………..27 X Figure 2.6: Mosquitoes gathering at the feeder and piercing through the membrane, containing HMBPP-supplemented gametocyte culture………………...30 Figure 2.7.a: Midgut of an An. funestus mosquito with a P. falciparum oocyst……31 Figure 2.7.b: Midgut of an uninfected An. funestus mosquito……………………….32 Chapter Three Figure 3.1: Agarose gel for the identification of members of the An. funestus Group…………………………………………………………………………33 Figure 3.2: Relative species composition of members from the An. funestus group (An. leesoni, An. vaneedeni, An. rivulorum) from the Mamfene region in KwaZulu-Natal as collected over a period of 8 months………………….34 Figure 3.3: Relative species composition of members from the An. gambiae complex (An. arabiensis, An. merus, An. quadriannulatus) from the Mamfene region in KwaZulu-Natal as collected over a period of 8 months………………………………………………………………………..36 Figure 3.4: The relative abundance of the major vector, An. arabiensis and minor and non-vector species from the An. funestus group and An. gambiae complex over an 8-month period. Average monthly rainfall (mm) is also indicated for the period…………………………….37 Figure 3.5: Optical density values for antibody detection as measured at 405 nM…………………………………………………………………………….39 Figure 3.6: Attraction of An. arabiensis females towards the control and treatment (HMBPP-supplemented blood) group…………………………41 XI Figure 3.7: Attraction of An. coluzzii females towards the control and treatment (HMBPP-supplemented blood) group…………………………41 Figure 3.8: Attraction of An. gambiae females towards the control and treatment (HMBPP-supplemented blood) group…………………………41 Figure 3.9: Attraction of An. funestus females towards the control and treatment (HMBPP-supplemented blood) group……………………….. 41 Figure 3.10: Attraction of An. merus females towards the control and treatment (HMBPP-supplemented blood) group…………………………42 Figure 3.11: Attraction of An. vaneedeni females towards the control and treatment (HMBPP-supplemented blood) group…………………………42 Figure 3.12: Attraction of An. quadriannulatus females towards the control and treatment (HMBPP-supplemented blood) group…………………………42 Figure 3.13: Dose response curve representing different HMBPP concentrations represented as percentage feeding rate. The feeding rates for both An. funestus and An. gambiae are illustrated…………………………….46 Figure. 3.14a & b: Percentage oocyst prevalence for the control and test group in An. arabiensis midguts (a) and average number of oocyst per midgut (intensity) in An. arabiensis females (b)…………………………….47 Figure. 3.15a & b. Percentage oocyst prevalence for the control and test group in An. gambiae s.s. midguts (a) and average number of oocyst per midgut (intensity) in An. gambiae s.s. females (b)…………………………48 Figure. 3.16a & b. Percentage oocyst prevalence for the control and test group in An. coluzzii s.s. midguts (a) and average number of oocyst per midgut (intensity) in An. gambiae s.s. females (b)…………………………49 XII Figure. 3.17a & b. Percentage oocyst prevalence for the control and test group in An. funestus s.s. midguts (a) and average number of oocyst per midgut (intensity) in An. funestus s.s. females (b)…………………………50 Figure. 3.18a & b. Percentage oocyst prevalence for the control and test group in An. merus midguts (a) and average number of oocyst per midgut (intensity) in An. merus females (b)…………………………………51 Figure. 3.19a & b. Percentage oocyst prevalence for the control and test group in An. quadriannulatus midguts (a) and average number of oocyst per midut (intensity) in An. quadriannulatus female (b)……..…………52 XIII LIST OF TABLES Chapter Two Table 2.1: Species-specific primer sequences for members of the An. funestus group…………………………………………………………..19 Chapter Three Table 3.1: The number of anopheline species from the An. funestus group collected over an 8-month period from sites in Mamfene, KwaZulu- Natal………………………………………………………………………..34 Table 3.2: The number of anopheline species from the An. gambiae complex collected over an 8-month period from sites in Mamfene, KwaZulu- Natal………………………………………………………………………..29 Table 3.3: Pearson’s correlation between the relative abundance of the major vector, An. arabiensis and all other minor- and non-vector species from the An. gambiae complex and the An. funestus group…………38 Table 3.4: Pearson’s correlation between rainfall and An. gambiae complex and An. funestus group members……………………………………….38 Table 3.5: Statistical analysis of the choice chamber evaluation with cow blood and sugar water as the two sources for attraction…………………….40 Table 3.6: Statistical analysis of the attraction towards HMBPP-supplemented blood (treatment group) and blood only (control group)………………43 Table 3.7: Statistical analysis of the feeding rate between HMBPP- supplemented blood (treatment) and blood only (control)…………...45 Table 3.8: Statistical significance between HMBPP-supplemented P. falciparum infected blood (treatment) and P. falciparum infected blood only (control) of An. arabiensis mosquitoes…………………………………………….47 XIV Table 3.9: Statistical significance between HMBPP-supplemented P. falciparum infected blood (treatment) and P. falciparum infected blood only (control) of An. gambiae s.s. mosquitoes………………………………………….48 Table 3.10: Statistical significance between HMBPP-supplemented P. falciparum infected blood (treatment) and P. falciparum infected blood only (control) An. coluzzii mosquitoes……………………………………………………49 Table 3.11: Statistical significance between HMBPP-supplemented P. falciparum infected blood (treatment) and P. falciparum infected blood only (control) of An. funestus s.s. mosquitoes………………………………………….50 Table 3.12: Statistical significance between HMBPP-supplemented P. falciparum infected blood and P. falciparum infected blood only (control) of An. merus mosquitoes…………………………………………………………………51 Table 3.13: Statistical significance between HMBPP-supplemented P. falciparum infected blood and P. falciparum infected blood only (control) of An. quadriannulatus mosquitoes…………………………………………52 Table 3.14. The overall effect of HMBPP on the attraction, feeding rate, oocyst prevalence and intensity of different anopheline mosquitoes……….53 XV ABBREVIATIONS SFMA: Standard Membrane Feeding Assay ELISA: Enzyme-linked Immunosorbent Assay PCR: Polymerase Chain Reaction HMBPP: (E)-4-hydroxy-3-methyl-but-2-enyl pyrophosphate - 1 - CHAPTER ONE: Introduction 1.1. Malaria Overview Malaria continues to be one of the most destructive diseases and despite efforts to fight it, an estimate of 229 million cases were reported in 2019 – of which approximately 409,000 were fatal (World malaria report, 2020). This infectious disease has shown to follow a seasonal trend with climate factors also contributing (Caminade et al. 2013). Furthermore, malaria endemic areas are more affected, since malaria transmission tends to present itself in short seasons or as epidemics that can result in high mortalities (Caminade et al. 2013). In an effort to eliminate malaria, the WHO introduced the global malaria eradication program in the 1950’s, during which 79 countries achieved elimination status. The most successful countries were however the ones with highly seasonal malaria transmission, including northern Africa and America as well as Australia, and Plasmodium vivax was mainly responsible for infections (Feachem et al. 2010). Limitations in public health services and the extend of the Plasmodium infections in the endemic areas of Africa resulted in the replacement of malaria elimination prospects with control policies by the WHO in the early 1970’s (WHO, 1970). To date the African region still accounts for approximately 94% of cases (World malaria report, 2020) and several countries are still affected by residual transmission. South Africa is among these countries where residual transmission can be attributed to the fact that it borders other African malaria endemic countries (Morris et al. 2013). Though it is known that malaria is caused by a protozoan parasite and that an Anopheles mosquito vector and a vertebrate host are needed (Laveran, 1884), the knowledge regarding malaria transmission still remains inadequate. It is known that about 60, out of more than 500 anopheline mosquito species, are described as malaria vectors (Gillies & De Meillon, 1968; Gillies & Coetzee, 1987; Manguin, 2008; Sinka et al. 2010) and 48 of approximately 250 Plasmodium parasite species are responsible for infecting mammal hosts (Faust & Dobson, 2015). Furthermore, the P. falciparum parasite remains the greatest contributor of this disease and it was responsible for 99,7% of estimated sub-Saharan Africa malaria cases in 2018 (World - 2 - Malaria Report, 2019). Its most efficient mosquito companions include the An. gambiae complex and the An. funestus group. The combination of the parasite and the latter species group and complex, are responsible for 90% of deaths in sub- Saharan Africa (Fontenille et al. 1997; Fontenille & Simard, 2004; Cohuet et al. 2004; Lwetoijera et al. 2014). Although there’s been a reduction of 40% in the prevalence of P. falciparum, across Africa, by controlling the mosquito vector with long lasting insecticide nets (LLINs) (Bhatt et al. 2015) malaria transmission still remains a problem. This could be explained by the fact that this control method might only encourage vector species to feed on people when they are not using LLINs (White et al. 1972; Mbogo et al. 1993; Kiszewski et al. 2004; Lyimo and Furguson, 2009; Takken & Verhulst, 2013). This includes people who don’t have LLINs as well as those being exposed during early mornings or evenings when not sleeping (Takken, 2002). Vector control strategies are further complicated by the fact that the Anopheles genus consists of species complexes and groups, where the different species within these complexes and groups exhibit different feeding behaviors and roles in malaria transmission (Gillies & Coetzee, 1987; Kyalo et al. 2017; Wiebe et al. 2017). The An. gambiae complex consists of nine sibling species which are morphologically indistinguishable and includes vector and non-vector species (Gillies & Coetzee, 1987; Barron et al. 2019; Coetzee 2020). Furthermore this complex comprises freshwater and saltwater breeders (Coetzee et al. 2013). The three major vectors of this complex include An. gambiae s.s., An. coluzzii and An. arabiensis which exhibit different behaviors. Anopheles gambiae s.s. preferentially bite human hosts indoors and An. coluzzii exhibit similar behavior, whereas An. arabiensis has a more diverse behavior and will feed indoors or outdoors on human and other mammalian hosts (Gillies & Coetzee, 1987; Sinka et al. 2010; Akogbéto et al, 2018). Anopheles funestus s.s. is the most anthropophilic and predominant member of the eleven members that form part of the African An. funestus group (Gillies & De Meillon, 1968; Gillies & Coetzee, 1987; Harbach, 2013). The latter species group also differs from the An. gambiae complex in terms of ecology, biology and genetics (Fontenille & Simard, 2004; Coetzee & Koekemoer, 2013). Consequently, depending on their role in malaria transmission, vectors can be classified as either major or minor (Cohuet 2009). Major malaria vectors are described as vectors that are responsible for more than 95% of the total malaria transmission globally and for sub-Saharan Africa these - 3 - species include: An. gambiae, An. arabiensis, An. coluzzii and An. funestus (Gillies and De Meillon, 1968; Gillies and Coetzee, 1987; Antonio-Nkondjio et al. 2006; Sinka et al. 2010; Coetzee et al. 2013; Kyalo et al. 2017). Minor vectors are thus responsible for the remaining malaria transmission, and these species include An. melas, An. merus and An. bwambae (from the An. gambiae complex) and An. rivulorum from the An. funestus group (Gillies & Coetzee 1987; Wilkes et al. 1996; Antonio-Nkondjio et al. 2006; Kawada et al. 2012; Kyalo et al. 2017). The interaction between the Plasmodium parasite and its vectors also plays an important role when considering malaria control interventions. These parasites, especially P. falciparum, have been found to alter the behavior of its mosquito vectors and consequently enhance malaria transmission (Cator et al. 2012). Recent studies found that the parasite metabolite, (E)-4-hydroxy-3-methyl-but-2-enyl pyrophosphate (HMBPP), is responsible for increased attraction and infection in a mixed strain mosquito vector of An. gambiae s.l. (Emami et al. 2017). This parasite- host manipulation phenomenon has however not been studied extensively in African malaria vectors and it is therefore important not to only concentrate on the mosquito vectors but to follow a holistic approach with regards to the transmission cycle between the parasite and the vectors. - 4 - 1.2 Aim, Objectives and hypotheses The aim of this study was to monitor the seasonal changes of vectors in KwaZulu- Natal and determine if mosquito species were affected by the presence of the P. falciparum metabolite, HMBPP. Furthermore, this study was conducted in order to gain important information concerning the complex interaction between the malaria parasite and its mosquito companion. As the parasite evolve to manipulate its vector, it is important to remain vigilant as human malaria infections might be affected as a result. Being informed about how different vector species behave in the presence of the parasite, could assist in following a holistic approach that might prevent further infection, for example by adjusting vector control strategies to target species that form part of the cycle. Specific objectives pertaining to the laboratory study included: ● Evaluating the seasonal change and overall mosquito vector dynamics between the major-, minor- and non-vectors of South Africa ● Determining the effect of HMBPP on the attraction of the major-, minor- and non-vector mosquito species ● Determining the effect of HMBPP on the feeding behavior of the major-, minor- and non-vectors ● Determining the effect of HMBPP on the susceptibility to P. falciparum of different vector species The main hypotheses included: 1. The vector dynamics of the minor- and non-vector species will change in the occurrence of changes in the major vector population 2. The feeding rate and attraction behavior will be altered in the presence of HMBPP 3. Since the blood meal size is suggested to increase in the presence of HMBPP (Emami et al. 2017), the prediction would be that more gametocytes would be ingested and thus a higher number of oocysts would be present in the mosquito midgut. - 5 - 1.3 Literature Review 1.3.1 Mosquito vector dynamics and seasonal variation of species The potential for an Anopheles mosquito to become a vector depends on several factors, including the ability to support malaria parasite development, environmental conditions, blood feeding behavior and adult longevity/survival (Beier, 1998). More recent studies provides support for the fact that environmental factors such as temperature and rainfall could affect the transmission potential of a vector, due to changes in the biology of the vector (Afrane et al. 2012; Paaijmans et al. 2009; Paaijmans et al 2012; Mordecai et al. 2013). Larval condition has also been found to have an effect on adult traits that are important for vector competence, such as adult longevity and body size (Plaistow et al. 2001; Pulkkinen & Ebert, 2004). Although the mechanisms responsible for regulating malaria parasite-vector interactions are still being evaluated, historically, it was proposed that only certain mosquito species are able to support the parasite cycle (Beier, 1998). According to Beier (1998) only certain mosquito species are able to take up gametocytes from parasite-infected blood, and support the formation of the sporozoite stage of the parasite, which are transmitted to humans. However, recent studies provided support for the hypothesis that most anophelines might be able to support parasite development, but due to host feeding preferences, it might limit their role in malaria transmission (De Meillon et al. 1977; Temu et al. 2007; Burke et al. 2017). These mosquito vector dynamics also play an important role, when considering control methods to reduce or eliminate certain mosquito species. It brings the question to mind: would these control methods, which focus on eradication or elimination of current major mosquito vector species, not create a niche for minor vector species to fulfill the roles of the current ones? Ecologists proposed the exclusion principle which is described as the existence of two non-interbreeding species in the same geographic area, where one species multiplies faster than the second species, and eventually replacement of the latter species will take place (Elton, 1927; Hardin, 1960). One form of species replacement takes place when introduced species numbers increase due to external interference, such as weather, habitat alteration or man-made intervention which in turn causes resident species numbers to decline (Herbold & Moyle, 1986). Species replacement was suspected to - 6 - occur in south west Uganda where the minor vector, An. rivulorum, replaced the major vector, An. funestus, due to the fact that the latter species was eliminated by house-spraying (Aboul-Nasr, 1970). The same phenomenan was observed in the inland parts of Kenya and Tanzania, where the elimination of An. funestus led to a sharp increase in the number of An. rivulorum mosquitoes (Gilies & Smith, 1960). It could therefore be suggested that if minor vector numbers are not limited (by low abundance or survival) and thus flourish, they can be considered as main vectors since these minor vectors might be able to maintain malaria transmission (Wilkes et al. 1996; Antonio-Nkondjio et al. 2006). This was also described by Temu et al. (1998) where the minor vector, An. merus, was found to be the main vector with the highest infection status. Additional studies by Quakyi et al. (2000) also proposed that high malaria transmission, which is maintained by the major vectors, could aid the minor vectors’ performance, by keeping the parasite cycle going. In addition to different malaria vector species and their role in malaria transmission, both transmission and parasitaemia might present seasonal variation to some extend (Smith et al. 1993). In such regions where low but periodic transmission occurs, epidemics might be the result (Bruce-Chwatt, 1985). Although the same transmission pattern is observed in South Africa (Hargreaves et al. 2000; Maharaj et al. 2013), in general malaria transmission is low and follows a seasonal trend in the malaria endemic provinces (Burke et al. 2019). Low-level transmission is described as transmission by minor vectors, of which members of both the An. funestus group and An. gambiae complex mosquitoes form part of (Gillies & Coetzee, 1987; Coetzee & Hunt, 1998). Low-level transmission could further be the result of effective malaria control that results in low abundance of main vector species. Furthermore, the current major vector species responsible for malaria transmission in South Africa includes An. arabiensis (a member of the An. gambiae complex) and historically it also included An. funestus (from the An. funestus group) (Hargreaves, et al. 2000; Brooke et al. 2013). Anopheles funestus s.s. is amongst the species that are responsible for malaria epidemics, when present in South Africa, (Fontenille et al. 1990; Hargreaves et al. 2000). Anopheles arabiensis is mostly responsible for the continuation of seasonal malaria transmission in the absence of An. funestus (Coetzee & Hunt 1998; Coetzee et al. 2013; Dandalo et al. 2017). Some minor vectors that are responsible for low- - 7 - level malaria transmission includes An. merus, a member of the An. gambiae complex, as well as An. vaneedeni and An. parensis (both members of the An. funestus group). Anopheles merus has been regarded as a minor vector in the country and has been directly involvement in malaria transmission in the southern parts of Mozambique (Cuamba & Mendis 2009). The fact that this species occurs in all the malaria endemic provinces of South Africa, is therefore a cause for concern (Mbokazi et al. 2018; Coetzee et al. 2000). Anopheles vaneedeni has been a suspected minor vector in sporadic malaria cases in the historic northern Transvaal (Gillies & Coetzee, 1987) and has recently been incriminated as a minor vector, as it is suspected to be involved in residual transmission in KwaZulu-Natal (KZN), South Africa (Burke et al. 2017). Another species that also recently tested positive for P. falciparum in the same province is An. parensis (Burke et al. 2019). Although several other members of the An. funestus group, including An. leesoni and An. rivulorum, are perennially present in KZN, their role in malaria transmission needs further investigation (Mouatcho et al. 2007; Burke et al. 2017). KwaZulu-Natal is amongst the South African malaria endemic provinces where the primary malaria control method is insecticide based indoor residual spraying (IRS). Anopheles arabiensis is currently the main vector in this province and relatively abundant in the study site (Dandalo et al. 2017), however, IRS does not target this species as efficiently as An. funestus due to the nature of its feeding and resting behaviours (Gillies & Coetzee, 1987; Sinka et al. 2010; Coetzee & Hunt 1998). The latter species was therefore controlled efficiently and could be eliminated, despite its up flare in the year 2000 (Hargreaves et al. 2000). A number of minor vector species co-exists with An. arabiensis, and also needs to be considered since control interventions, combined with seasonal variation, might enhance the transmission potential of minor vectors. Furthermore, as vector dynamics are constantly changing and more minor vector species are being incriminated in malaria transmission, it would be suspected to see an increase in the abundance and parasite transmission capacity of current minor vector species, as major vector numbers are declining, due to external factors such as control interventions. This phenomenon was also visible in the Kruger National Park where - 8 - An. arabiensis numbers declined due to possible draught in the area which resulted in an increase in the number of An. quadriannulatus (a non-vector species) mosquitoes (Munhenga et al. 2014; Munhenga, unpublished data). 1.3.2 Mosquito blood feeding and attraction behavior The blood sucking habit of insects is believed to have originated from a close and extended association with vertebrates (Lehane, 2005). Even unspecialized insects adapted to become blood-sucking ones as a result of the high nutritive value and easy digestibility of blood (Lehane, 2005). The estimated number of blood feeding insect species is 14,000 of which only 300 – 400 are of medical and agricultural importance (Adams, 1999; Lehane, 2005). Numerous blood-sucking insects have caused huge destruction as several diseases are caused by the transmission of parasites to both humans and animals (Lehane, 2005). Like for most other blood- sucking insects, mosquitoes are dependent on blood for the vital nutrients and proteins that are needed for egg production and therefore blood feeding is crucial for obtaining reproductive fitness (Takken & Verhulst, 2013). The term blood feeding can be described as a combination of behaviors which includes host detection, probing and piercing, location and ingestion of blood as well as termination of the feed (Friend & Smith, 1977). It is assumed that the blood source selected by female mosquitoes depends on genetic factors (Kilpatrick et al. 2007; Huang et al. 2009) and Southwood (1977) also proposed that the historic enviro nMent of a species influences its niche. All insects do not feed equally well from the same host and thus exhibit host choice preference. Host preference is a behavioral trait that is believed to have been shaped by native and external factors (Takken & Verhulst, 2013). Mosquitoes can be classified as either specialist feeders, which tend to feed on certain hosts, whereas the generalists are more opportunistic and make use of a variety of hosts (Takken & Verhulst, 2013). Furthermore, it was proposed that niche shifts and adaptation to different resources is possible (Chaves et al. 2010). This is due to the holometabolic trait of mosquitoes, which allows them to exploit different resource i.e. feeding on different hosts (Chaves et al. 2010). Feeding behavior also differs between different mosquito vector species. Anopheles arabiensis tends to rest outdoors and in the absence of insecticides, it will - 9 - also rest indoors together with exhibiting a diverse feeding habit (both human and cattle), whereas An. funestus is a highly anthropophilic and endophilic mosquito and thus mostly feed on human hosts (Coetzee & Fontenille, 2004; Sinka et al. 2010). In order for a mosquito to feed on a host, the host must be located and although some of the blood feeding mosquitoes are not fastidious when it comes to finding the right host (Takken & Verhulst, 2013), host attraction plays an important role in the location process. Several factors need to be taken into account when considering mosquito-host attraction, including heat, visual cues and olfaction (Takken & Knols, 1999; Lehane, 2005). At close proximity, mosquitoes can sense heat with thermosentitive neurons that are found in the sensilla on the tip of the antennae (Khan et al. 1966; Gingl et al. 2005). (Fig. 1.1). Visual cues, on the other hand, are perceived by the compound eyes of the mosquito, which can detect differences in light intensity (Muir et al. 1992; Moon et al. 2014). Odorants that are emitted by the host, are received by the olfactory receptors, which are located within sensilla on the palps and antennae of the mosquito (Fig.1.1) (Meijerink et al. 2001; Qiu et al. 2006; Carey et al. 2010). Some of these cues are host blood source, human skin bacteria and carbon dioxide (CO2) (Dekker et al. 2005; Verhulst et al. 2011). Despite the fact that some colonized An. gambiae strains lack effective CO2 receptors, these mosquitoes are still able to locate a host (McMeniman et al. 2014). Furthermore, the skin bacteria are responsible for producing volatile compounds, such as lactic acid, ammonia and ketones (Davis & Sokolove, 1976; Braks et al. 2001; Verhulst et al. 2010). Body heat and relative humidity form part of external cues that are responsible for the dispersal of semiochemicals by convection currents which are generated by body heat and humidity (De Jong & Knols, 1995; Olanga et al. 2010). Olfactory cues are received by chemosensory receptors, including odorant receptors, ionotropic receptors and gustatory receptors (Suh et al. 2014). These receptors allow mosquitoes to integrate taste and odor in the host location process (Riabinina et al. 2016). The orco gene that is expressed by the olfactory receptor neurons, connects the suboesophageal zone of the brain to the labella on the proboscis, in order to facilitate this process (Riabinina et al. 2016). - 10 - Fig.1.1. Head and mouthparts of a female anopheline mosquito. 1.3.3 Parasite host manipulation The presence of parasites is another important factor that needs to be considered as studies have shown that mosquito feeding behavior and attraction is affected upon infection of the mosquito with human malaria parasites (Koella & Packer, 1996; Anderson et al. 1999; Emami et al. 2017). Parasites that increase malaria transmission by changing the behaviour of their hosts, is not a scarce phenomenon in nature (Moore, 2002). Though several studies have suggested that parasites alter the feeding behaviour and thus increase the feeding rate of the vector, once the mosquito is already infected (Rossignol et al. 1984; Schwartz & Koella, 2001; Koella et al. 2002), others found that infection status did not determine feeding behaviour (Li et al. 1992; Hogg & Hurd, 1995). In addition to changes in feeding behaviour, the P. falciparum parasite has also been found to enhance attraction of its vectors to the infected hosts. The P. falciparum metabolite, HMBPP, has been suggested to be responsible for these manipulative changes in the mosquito (Emami et al. 2017). Mosquitoes’ parasite companions (in particular P. falciparum and P. vivax) are host specific (Takken & Verhulst, 2013) and therefore these parasites would guide the mosquitoes to find a human host. The ‘manipulation hypothesis’ has been described as the way in which the malaria parasite enhances transmission by adaptive Antennae Palps - 11 - manipulation of the malaria vector (Cator et al. 2012). Evidence suggested that manipulation by the parasite, by means of increasing the mosquito feeding rate, together with causing changes in the probing behavior of sporozoite infected females, might change the biting frequency of the vector (Cator et al. 2012). This in turn also increases the likelihood of the vector to attempt to feed and thus enhance possible transmission (Cator et al. 2012). This was also supported by a study where An. gambiae mosquitoes that were infected with P. falciparum sporozoites, showed elongated probing durations and an increase in the number of probes while feeding (Wekesa et al. 1992). As a result, the chance of more sporozoites being transmitted to the human host increases and therefore enhanced transmission of the parasite is facilitated through this probing behavior (Wekesa et al. 1992). It has also been suggested that the parasite is responsible for possible changes in mosquito behavior, through the manipulation of host odors (De Moraes et al. 2014). This might result in a cycle where the infected human hosts are more attractive to the mosquitoes, resulting in higher anthropophilic feeding behavior, which will dramatically influence malaria transmission (Takken & Verhulst, 2013). A study in Kenya showed that children with P. falciparum gametocytes were more attractive to mosquitoes than those without gametocytes (Lacroix et al. 2005; Busula et al. 2017). An explanation for this is that attractive volatile compounds such as terpenes that are produced by malaria parasites result in increased attraction of mosquitoes to the host (Nyasembe et al. 2012; De Moraes et al. 2014; Kelly et al. 2015, de Boer et al. 2017). Studies conducted in human, mouse and avian malaria parasites further suggested that volatile organic compounds (VOCs) found in exhaled breath, are changed by Plasmodium infection and this might consequently result in increased recognition by the mosquitoes (Schaber et al. 2018). Although human odor, which consists of numerous chemicals, play a critical role in indicating human presence (Bernier et al. 2000; Takken et al. 1999), P. falciparum terpenes also play a vital role in assisting with mosquito attraction. These parasites produced terpenes, which form part of the attractive volatile compounds, and fall into the isoprenoids class (Emami et al. 2017). Furthermore, these terpenes consist of isopentenyl (IPP) and its isoform dimethylallyl pyrophosphate (DMAPP) (Fig.1.1). The synthesis of IPP and DMAPP occurs through the mevalonate pathway that is used by humans, mosquitoes and other higher eukaryotes (Emami et al. 2017). This crucial - 12 - metabolic pathway is involved in several cellular processes by the synthesis of isoprenoids – both sterol (like cholesterol) and non-sterol (Buhaescu & Izzedine, 2007). In contrast to this, apicomplexan parasites, such as P. falciparum, and other eubacteria use an alternative pathway, known as the 2-C-methyl-D-erythritol 4- phosphate (MEP) pathway (Fig.1.1), for IPP and DMAPP synthesis (Morita et al. 2007; Emami et al. 2017). The MEP pathway uses (E)-4-hydroxy-3-methyl-but-2-enyl pyrophosphate (HMBPP) as a precursor for the production of these isoprenoids (IPP and DMAPP) (Emami et al. 2017). HMBPP is not only known to activate human Vɣ9Vδ2 T-cells (which was found to trigger immune response in An. gambiae s.l.) (Morita et al. 2007; Lindberg 2013), but has also been found to trigger red blood cells to release VOCs. This is based on the same principle in which P. falciparum infected red blood cells act as an attractant to stimulate blood feeding by mosquito vectors (Emami et al. 2017). There is however insufficient knowledge regarding the effect of HMBPP on African mosquito species. - 13 - Mevalonate pathway MEP pathway Acetyl-CoA Pyruvate + glyceraldehyde 3-phosphate HMG-CoA 1-Deoxy-D-xylulose 5-phosphate Mevalonate 2-C-methyl-D-erythritol 4-phosphate Mevalonate-5-phosphate 4-Diphosphocytidyl-2-C-methylerythritol Mevalonate-5-pyrophosphate 4-Diphosphocytidyl-2-C-methylerythritol 2-phosphate 2-C-methyl-D-erythritol 2,4-cyclodiphosphate HMBPP (E)-4-hydroxy-3-menthyl-but-enyl-pyrophosphate IPP DMAPP Isopentenyl pyrophosphate Dimethylallyl pyrophosphate Fig. 1.2. Isoprenoid biosynthesis through the mevalonate and MEP pathways. The mevalonate pathway is found in higher eukaryotes whereas the MEP pathway is used by most eubacteria and apicomplexan protozoa. OPP OPP - 14 - 1.3.4 Parasite susceptibility of different mosquito vectors Gametocytes and sporozoites are the big culprits in the transmission cycle between the parasite and the vector (Delves et al. 2013). The female and male gametocytes form part of the sexual stage of the parasite and are ingested by the mosquito when taking a blood meal from an infected human host (Smith et al. 2014). The sexual phase of P. falciparum needs to develop in vivo and can be separated into gametocytogenesis and gametogenesis (Sinden, 1981). Gametocytogenesis is the phase that takes place within the mosquito vector and is necessary for the mature gametocytes to differentiate into motile male microgametes and a single sessile female macrogamete (Sinden, 1981). Gametogenesis, on the other hand, involves the activation and fertilization preparation of gametocytes (Bousema & Drakeley 2011). Exflagellation is the term that is used to describe the transformation process of male gametocytes into gametes as well as their emergence from red blood cells (Delves, 2013). Since the exflagellation process takes place within the mosquito, several suggestions have been made to further explain vector susceptibility to the parasite. Species-specific stimulation of exflagellation was earlier considered a possible mechanism for aiding mosquito susceptibility (Micks et al. 1948; Nijhout, 1979). Recent studies found that one of the important transformation processes (gametocytes into gametes), that forms part of the development of the P. falciparum parasite, is triggered by a favourable change in enviro nMent within the mosquito gut. Among these favourable conditions is the presence of the mosquito derived gametocyte-activating factor xanthurenic acid (Garcia et al. 1998). The latter is among the contributing factors, which enables gametocytes to become mature and thus enhance the chance that a mosquito becomes infective (Oaks et al. 1991). Other factors include oxygen and carbon dioxide concentration, pH and temperature factor (Oaks et al. 1991). After exflagellation, the fusion of the male and female gametes takes place as well as the formation of a zygote from which the motile ookinete originates (Sinden, 1981). The ookinete exits the bloodmeal, which was taken up by the mosquito, and it enters the midgut epithelium of the mosquito (Sinden, 1981). The ookinete matures and - 15 - forms an oocyst from which motile sporozoites originate (Oaks et al. 1991). These sporozoites travel to the salivary glands of the mosquito, from where they are injected into the host upon being bitten by the mosquito (Oaks et al. 1991). In order for the parasite to become infective, it needs to undergo an incubation period of 14- 15 days (MacDonald, 1955; Emami et al. 2017) Vector susceptibility is an important parameter for parasite transmission (Cohuet et al. 2009). A combination of a vector’s susceptibility and parasite infectivity could be described as vector competence and it involves both the host’s defence mechanisms and the parasite’s ability to overcome those defences (Lefevre, 2013). The susceptibility of certain Anopheles species to the malaria parasite is a mysterious phenomenon and the mechanisms involved is not well known (Oaks et al. 1991). Weathersby (1952) suspected that the inability of certain mosquito species to sustain the development of the parasite could be due to a toxin being present that inhibits the development of the parasite. It might also be that the absence of factors, which are essential for parasite development, plays a role (Weathersby, 1952). More recent studies have shown that nitric oxide (NO) is also involved in termination of parasite development. Once the mosquito ingests infected blood from a human host, there is a substantial increase in the levels of nitrates and NO in the mosquito midgut (Luckhart et al. 2003; Peterson et al. 2007). The midgut epithelium could also naturally contain NO and upregulation of NOs has also been detected in An. gambiae midguts (Dimopoulos et al. 1998; Luckhart et al. 1998). Oxyhaemoglobin and haeme which are released by the digestion of red blood cells in the mosquito, can react with NO to produce NO metabolites which are toxic and can negatively influence parasite development (Peterson et al. 2007). Besides vertebrate-derived components of blood that forms part of the modulatory components in preventing parasite development, mosquito immunity also involves defence peptides which can inhibit sporogonic development (Sinden et al. 2004). The development of proteins on the surface of the parasite, once inside the midgut of the mosquito, could also be recognized by the mosquito antibodies and as a result inhibit further development of the parasite (Gozar et al. 1998; Margos et al. 1998). In addition to this, melanisation and lysis of Plasmodium parasites also occurs. A strain of An. gambiae has been found to melanise the ookinetes of P. berghei and allopatric strains of P. falciparum whereas another strain of An. gambiae lysed P. gallinaceum - 16 - ookinetes (Collins et al. 1986). Melanisation of these ookinetes occurs as they emerge from the midgut epithelium whereas the lysis of ookinetes takes place while these ookinetes traverse the midgut epithelium (Vernick et al. 1995). Furthermore, for mosquitoes to be classified as successful vectors of the parasite, they need to carry sporozoites to their hosts. If these sporozoites are not able to migrate and successfully penetrate the mosquito salivary glands, it might be another possible factor contributing to the mosquito’s resistance to parasites (Collins et al. 1986). Measurable infection related traits could be evaluated in order to determine the competence of a vector species – these are parasite prevalence and intensity (Lafevre, 2013). Prevalence could be described as the proportion of mosquitoes that have been exposed to the parasite, which harbours at least one oocyst in the midgut, whereas parasite intensity denotes the oocyst number per infected gut (Lafevre, 2013). Both low prevalence and intensity could indicate high anti-infection resistance by the mosquito or high antigrowth resistance to parasite reproduction in the mosquito, respectively. High prevalence and intensity would therefore imply the opposite (Lafevre, 2013). HMBPP was found to be released by P. falciparum into the blood and to regulate sporogenesis in the mosquito (Emami et al. 2019). Furthermore, it was shown to increase the size of the blood meal taken by the mosquito and there by increases the number of oocysts in the mosquito midgut (Emami et al 2019). The addition of HMBPP to P. falciparum blood would therefore increase the likelihood of higher oocyst prevalence and intensity in African mosquito vectors as well. - 17 - CHAPTER TWO: Materials and Methods 2.1 Sampling of wild mosquitoes The wild mosquitoes in this study were collected at sentinel sites which were established by members of the Sterile Insect Technique (SIT) project. These sites were located in Mamfene, in the Umkhanyakude district of KwaZulu-Natal (KZN). Collections varied between larval collections and adult mosquito collections (with mouth aspirators). Resting adult mosquitoes were mainly collected from clay pots (Fig. 2.1a) or a modified plastic bucket (Dandalo et al. 2017). In addition, some adults were collected from tyres around cattle kraals (Fig.2.1b) as well as CO2 traps which were placed at sentinel sites ad hoc during this time. Collections ranged over an 8- month period (Aug. 2019 – March 2020) as An. funestus group species were not available during the months prior to this time and due to COVID 19 restrictions, no fieldwork was conducted after March 2020. Mosquitoes were collected monthly and apart from assisting with fieldwork occasionally, mosquitoes were mostly collected by the SIT team based in KZN. Adult mosquito collections, as well as the larvae and pupae, were brought back to the Vector Control Reference Laboratory and the Wits Research Institute for Malaria in Johannesburg, South Africa, where further rearing and processing of samples took place. Adult female mosquitoes were placed into egg laying vials (Choi et al. 2014), of which each contained a piece of wet filter paper to provide the appropriate substance for egg laying. The F1 generation was further reared through to adults, which were then preserved in silica tubes for further morphological and molecular identification. - 18 - Fig. 2.1a &b. Resting adult mosquitoes being collected from a clay pot and the search for resting adult mosquitoes in tyres at a sentinel site in Mamfene, KwaZulu-Natal. 2.2. Wild caught mosquito identification Mosquito identification needs to be conducted on both morphological and molecular level, in order to ensure accurate species identification (Erlank et al. 2018). Mosquitoes were separated, based on morphological differences, with the use of a dichotomous key by Coetzee (2020). Adult mosquitoes were morphologically separated into Anopheles gambiae complex, An. funestus group and other anophelines. After the morphological identification, further molecular identification was conducted by the polymerase chain reaction (PCR) assay. As An. gambiae complex species data was generated for the SIT project, PCR assay was conducted by the relevant team and did not form part of this study. 2.2.1 PCR species identification and seasonal trend For identification of An. funestus group mosquitoes on a molecular level, DNA was extracted from a mosquito leg or wing of this species group with the use of the prepGEM Insect DNA extraction kit (ZyGEM; Cat no. E3BG6M; New Zealand). An insectary reared An. funestus mosquito, from the FUMOZ colony, was used as a positive control and a negative control was included during the DNA extraction assay and PCR assay. Post DNA extraction, specimens were further analysed with the use - 19 - of species-specific polymerase chain (PCR) reaction assay according to the PCR protocol for the An. funestus group (Koekemoer et al. 2002; Cohuet et al. 2003). A PCR master mix was made up, consisting of 1.25 μl of 10 x reaction buffer (100 mM Tris-HCL, 500 mM KCl), dNTP mixture (which includes 2.5 mM of each dNTP), 25 mM MgCl2 (ThermoFisher Scientific; Cat no. A31021; South Africa). A concentration of 3.3 pmol of each of the following primers: universal, An. funestus, An. leesoni, An. parensis, An. rivulorum, An. rivulorum-like and An. vaneedeni (primer sequences indicated in Table 2.1) was added. This was followed by the addition of Taq- polymerase (5 units/µl) (ThermoFisher Scientific; Cat no. R001AM; South Africa) and 3.65 µl deionised H2O. Each volume was multiplied by the number of samples to be analysed, including the positive control and negative controls for DNA extraction and the PCR master mix, and the mixture was vortexed and centrifuged. Aliquots of 12.5 µl was then made into the individual PCR tubes, containing 0.5 µl of the sample DNA template. The PCR tubes were placed into the PCR machine (Biorad T100 Thermal Cycler) which was programmed according to the following cycling conditions: 94 ̊C for 2 minutes and a 30 x repetition of the following conditions: 94 ̊C for 30 seconds, 45 ̊C for 30 seconds, 72 ̊C for 40 seconds and 72 ̊C for 5 minutes. Table 2.1: Species-specific primer sequences for members of the An. funestus group (Koekemoer et al. 2002; Cohuet et al. 2003). Primer name Sequence (5’ to 3’) UV TGT GAA CTG CAG GAC ACA T FUN GCA TCG ATG GGT TAA TCA TG VAN TGT CGA CTT GGT AGC CGA AC RIV CAA GCC GTT CGA CCC TGA TT RIVLIKE CCG CCT CCC GTG GAG TGG GGG PAR TGC GGT CCC AAG CTA GGT TC LEES TAC ACG GGC GCC ATG TAG TT The PCR product of each assay was electrophoresed on a 2.5% Tris–acetate– ethylenediamine tetraacetic acid (TAE) agarose gel, which was stained with ethidium bromide (0.5 μg/ml). Gels were viewed with a Geldoc system where after amplicon sizes were compared against the corresponding molecular weight marker: An. vaneedeni (587 bp), An. funestus (505 bp), An. rivulorum (411 bp), An. rivulorum-like (313 bp), An. parensis (252 bp) and An. leesoni (146 bp). - 20 - Species identification results for the An. gambiae complex mosquitoes were obtained from the SIT team. This was included in the results section in order to compare the abundance of the current major vector in the area, An. arabiensis (which is a member of the An. gambiae complex), to the minor and non-vector numbers of the An. funestus group. Monthly rainfall data were also obtained from the weather station in Mamfene in order to establish if there is a correlation between mosquito abundance and rainfall during this period. No temperature or humidity data was however available due to these sensors being damaged by cattle. 2.2.2 Data analysis The Pearson’s correlation analysis was used in order to determine if there was a positive correlation between the abundance of species within the An. funestus group and An. gambiae complex as well as between the major vector, An. arabiensis and minor- and non-vector species of both the An. gambiae complex and An. funestus groups. Pearson’s correlation was also used for rainfall and mosquito abundance. Analysis was conducted in GraphPad Prism (9.0.1) (Motulsky, 1998) and results were displayed as the r-value which is indicative of the coefficient of determination. A two-tailed P-value indicates the goodness of fit where a significant P-value rejects the idea that any correlation is due to random sampling. 2.3. Wild caught mosquito infection profiling with the Enzyme linked Immunosorbent Assay The infectivity of wild mosquitoes was detected by the Enzyme linked Immunosorbent Assay or better known as ELISA Wittz et al. 1987). This immunoassay is based on the principle where the enzyme acts as a signal generator in conjunction with a solid phase reactant and thus indicates an anitigen-anitibody reaction which is signalled through an enzyme (Beier et al. 1987). This assay has been widely used in antibody detection across human and animal illnesses and for the detection of the P. falciparum parasite, a repetitive antigen on the - 21 - circumsporozoite protein of the latter parasite is recognized by a monoclonal antibody as used in the ELISA assay (Beier et al. 1987). The P. falciparum ELISA assay was conducted as per the Malaria Research & Reference Reagent Resource Center (MR4) protocol over a period of two consecutive days. The P. falciparum sporozoite ELISA Reagent Kit (BEI Resources: https://www.beiresources.org/, NIAID, NIH: Plasmodium falciparum Sporozoite ELISA Reagent Kit, contributed by Robert A. Wirtz, Cat No. MR4-890; United States) was used. The following buffer solutions were prepared prior to analysis. A stock solution of phenol red (100 mg/ml) (ThermoFisher Scientific; Cat No. ICN10260405; South Africa) was prepared. A 1 x Phosphate Buffered Saline (PBS) solution was also made by dissolving PBS tablets (ThermoFisher Scientific; Cat No. BP2944; South Africa) in water and maintaining a concentration of 10 mM and pH of 7.4. Blocking buffer consisted of 2.5g casein (Merch; Catalog No. C7078; South Africa), 0.1 N NaOH (ThermoThermoFisher Scientific; Cat No. AC124260010; South Africa) and 10 mM 1x PBS solution. The 0.1 N NaOH was boiled in a flask and casein was slowly added. Once the casein dissolved and cooled down to room temperature, the PBS was added. A 1N concentration of HCL was added in order to reach a pH of 7.4. Phenol red (100 mg/ml) was then added to the blocking buffer (BB) to act as pH indicator. A 1x PBS- Tween solution was prepared by mixing 10mM PBS with 0.05% Tween-20 (ThermoThermoFisher Scientific; Cat No. BP337; South Africa). Lastly the grinding (homogenising) buffer was prepared by adding 125 µl of Octylphenoxy poly(ethyleneoxy) ethanol (IGEPAL CA-630) (ThermoThermoFisher Scientific; Cat No. ICN19859650; South Africa) to 25 ml of BB. A ratio of 1:1 of solution A to solution B (Qiagen; Cat No. 51200032; South Africa) was achieved in order to prepare the 2, 2’-azino-di (3-ethylbenzthiazoline-6-sulfonate) (ABTS) substrate solution. This latter mixture was wrapped in foil due to its photo-sensitivity. All the solutions were vortexed and stored at 4 ̊C for 1 week. A 96-well clear, round bottom PVC microtitre plate (Merck (Sigma-Aldrich); CAS number 2797, South Africa) was coated with the Plasmodium falciparum monoclonal antibody (mAb) Pf2A10 (BEI Resources, https://www.beiresources.org/) (0.5 mg/ml)) of which the stock solution consisted of mAb and a 1:1 glycerol to water solution. A - 22 - volume of 50 µl of the solution was added to each well and the plate was covered and incubate for 24 hours at 4 C̊. Mosquitoes (including 7 unfed female mosquitoes, from the FUMOZ colony, that were used as negative controls) were prepared by dissecting the mosquito and removing the head and thorax from the abdomen, with the use of a scalpel. The same method was used for the wild caught mosquitoes, which were preserved individually in 1.5 ml microcentrifuge tubes with silica gel, after collection. Each sample was removed from the tube and following the removal of the head and thorax, each sample (together with the negative controls) was placed into an individual 1.5 ml microcentrifuge tube with 50 µl grinding buffer. A 3 mm stainless steel bead (BMG Spartan, South Africa) was added, where after tubes were placed in the TissueLyser II (Qiagen) and were shaken at a high speed for 4 minutes at a frequency of 25Hz for 4 minutes. After samples homogenised, each bead was removed and 150µl of blocking buffer was added to each tube, to have a total volume of 200µl. Samples were stored at 4 ̊C for 24 hours and after usage remaining samples were stored at -20 ̊C for periods longer than 24 hours. After the 24 hour incubation period of the P. falciparum antibody coated plate, PSB- antibody mixture was aspired from the plate wells. The wells were filled with blocking buffer and left to incubate for 1 hour at room temperature. After the incubation period, the blocking buffer was poured off and samples and controls were added to the plate. The positive control, a purified recombinant protein, was prepared by adding 20 µg of capture mAb stock (Wirtz et al. 1987, BEI Resources, https://www.beiresources.org/) to 5 ml of PBS. The positive control was added to well A1 and 50 µl of each negative control was added to the last seven wells. Samples were added in the same quantity to the remainder of the wells and the sample number was recorded against the corresponding well number. The plate was covered with foil and incubated for two hours at room temperature. After incubation the wells were aspired and the plate was washed twice with PBS- Tween. A volume of 50 µl PfA10 peroxidase labelled antibody (0.5 mg/ml) diluted in BB (for a final concentration of 5.0 μg/5 ml) was added to each well. Due to the photosensitivity of the peroxidase labelled antibody, the plate was covered with foil - 23 - and incubated at room temperature for an hour. After the incubation period, the wells were aspired and the plate was washed four times with PBS-Tween. A volume of 100 µl of ABTS substrate solution was added to each well and the plate was incubated for 30-60 minutes to allow for the colour change to take place. The absorbency was determined for each well at 405 nM, with the use of a plate reader (Labsystems Multiskan RC) and Ascent software version 2.6. The cut-off value was determined with the following calculation: 2 x mean absorbance values of negative samples. Samples with absorbance values above that of the cut-off value, was regarded as positive and those with absorbance values below the calculated cut- off was regarded as negative. Positive samples had to be repeated by first boiling the sample at 100°C for 10 minutes and repeating the process in order to confirm the presence of the circumsporozoite protein (Durnez et al. 2011). 2.4. Attraction and feeding rate 2.4.1 Biological material Colonised mosquitoes: Insectary mosquito colonies that were used during this study included: FUMOZ (An. funestus from Mozambique, colonized in 2000 by Hunt et al., (2005)), KWAG (An. arabiensis from Mamfene, KwaZulu-Natal, SA, colonized in 2005 by Mouatcho et al. (2009), COGS (An. coluzzii from the Congo, colonized in 2009, (Koekemoer et al. 2011), MAFUS (An. merus from the Kruger National Park, SA, colonized in 2012 (Munhenga et al. 2014)), SANGWE (An. quadriannulatus from Zimbabwe, colonized in 1998). Mosquitoes were reared in a biosafety level 2 (BSL2) insectary and conditions were kept at 25 ̊C and a relative humidity of 80% (Hunt et al. 2005). The lighting cycle was set at 12-h day/night, including dusk/dawn cycles that lasted 45 minutes. Adult mosquitoes were maintained on a 10% sucrose solution. Wild mosquitoes: Mosquitoes used for the feeding or attraction study, were the progeny of the females (section 2.1) collected and placed in egg laying vials (Fig. 2.2.). Due to the low number of eggs that were produced by wild female mosquitoes, the F1 generation was very limited for this study, therefore on average 15 mosquitoes were used for each of the three technical replicates. All adult mosquitoes - 24 - were identified with the use of the dichotomous key of Coetzee (2020) and species identification was conducted by PCR assays according to the standard protocols for the An. gambiae complex (Scott et al. 1993) and the An. funestus group (Koekemoer et al. 2002) as described in section 2.2. Fig. 2.2. Egg laying vial containing a wild caught female Anopheles mosquito 2.4.2. Mosquito attraction and dual choice chamber assays i. Evaluation of the choice chamber Prior to conducting attraction assays, as described in the next section, the effectiveness of the choice chamber (Fig. 2.5.a) was evaluated using colonized mosquitoes An. funestus s.s. (FUMOZ) and An. arabiensis (KWAG). Experiments were conducted separately on each species and included three biological replicates of 20 female mosquitoes per replicate. The mosquitoes were provided with a choice between sugar water and cow blood (collected from the Karan beef abattoir as part of the larger SIT project). Each source was located at a different arm of the chamber. The blood and sugar water sources were also rotated between the two arms, with each biological replicate to ensure that positioning effects do not interfere with the source that was selected by the female (Fig. 2.5.b & c). As the laboratory mosquito - 25 - colonies were sustained on sugar water daily, the blood would have been the expected source of preference. Analyses were conducted and described in more detail (section iii) below. An additional control measure, to assess if mosquitoes alternate between the two arms (or between different food sources), was not evaluated at this stage. ii. Attraction behavior of major-, minor and non-vectors in the presence of HMBPP The dual choice chamber (Fig.2.3.a) was used to evaluate the attractiveness of mosquitoes towards human blood only (control group) vs. HMBPP-supplemented 15 cm 10 cm Main chamber Mosquitoes entering chamber Fig. 2.3.a, b & c. Dual choice chamber with dimensions of the main chamber and two arms (a) Dual choice chamber evaluation with sugar water and cow blood (b) Mosquitoes entering the preferred arm of the choice chamber (c). a b c 10 cm Perspex slides to section off arms post evaluation - 26 - human blood (test group). Species included all four major African vectors (An. gambiae s.s., An. coluzzii, An. arabiensis and An. funestus s.s.) as well as the minor vector, An. merus, and non-vector, An. quadriannulatus obtained from colonies. Wild caught An. vaneedeni, member of the An. funestus group, was also included in this study. Two glass feeders were set up according to the standard membrane feeding assay (SMFA) protocol as described in section 2.4.3. Blood type O+ was used and was loaded into the feeders - one feeder was used for the test group which consisted of 10 µM HMBPP (Merck (Sigma-Aldrich); CAS number 396726-03-7; South Africa) that was added to 1 ml blood. The feeder, containing HMBPP-supplemented blood, was placed on top of the net that covered the one arm of the choice chamber, whereas a different feeder was placed on the other arm, for the control group (blood only). With each biological replicate, the sources were alternated between the left and right arm of the chamber. Mosquitoes were starved for 24 hours (with distilled water provided), prior to feeding. Females (5 - 7 day old) were transferred into the main arm of the chamber one-by-one. Each mosquito was allowed 1 minute to choose which arm to enter, before the next mosquito was blown into the chamber. Mosquitoes were free to alternate between the two arms as well as to move around in the main section of the chamber. An additional control measure, to assess if mosquitoes alternate between the two arms (or between different food sources), was not evaluated at this stage. After 20 minutes, the arms were isolated from the main chamber by closing the gates (Fig 2.3.b. & c). The total number of mosquitoes in each arm section were recorded. Three biological replicates of 20 mosquitoes per species were conducted. The percentage of mosquitoes selecting the specific blood source was calculated and the preference index was calculated in R-Studio (1.3.1093) with the generalized linear mixed model (GLMM) as described in the next section. iii. Data analyses To account for variance between biological datasets, the generalized linear mixed model (GLMM) was used (Crawley, 2013) with the multiple regression model (lme4 package) incorporating replicate as a random factor for both control and treatment groups. The use of the GLMM models with random variables do not restrict statistical evaluation to normal distributed data. Thus, semi-parametric models are used for - 27 - evaluating attraction and feeding behaviour during this study. All analyses were done in R-Studio and results were entered into GraphPad Prism (9.0.1) to construct the graphs. 2.4.3. Mosquito feeding rate and HMBPP In order to determine if HMBPP plays a role in mosquito feeding rate, the SMFA was conducted and the feeding rate was determined between the control (blood) and treatment (HMBPP-supplemented blood) groups for different mosquito species. The SMFA was conducted by supplying blood through a glass feeder system that was connected to a water bath of which the temperature was maintained at 37C (Fig. 2.4). One millilitre of blood (O+ human blood that was donated) (Ethics no. M200692, Appendix B) was added to the glass feeder, which was covered with cow intestine at the bottom to form a membrane, of the control group. The same applied to the test group but in addition to adding 1 ml blood to the feeder, 10 µM of HMBPP was added and mixed with the blood, before adding it to the feeder. Since a concentration of 10 µM/ml is similar to the concentration of undiluted gametocyte- infected red blood cells (Emami et al. 2017), the concentration was kept constant throughout the study. Each blood filled glass feeder was placed horizontally on top of each cup (350 ml). Each cup consisted of 25 unfed female An. gambiae complex or 30 An. funestus group mosquitoes that were allowed to feed for 40 minutes. Prior to conducting the SMFA, unfed females were starved for 24 hours with distilled water soaked cotton pads. Each replicate consisted of three technical replicates of both the control and treatment groups and three biological replicates were conducted. After feeding, mosquitoes were immobilised by placing them in the freezer for 10 minutes and the number of unfed (Fig. 2.5.a) versus fed (Fig. 2.5.b) mosquitoes was recorded for each treatment. - 28 - Fig. 2.4. Set-up for the standard membrane feeding assay with feeding cups placed under the feeding system with water (37 ̊C) circulating from the water bath (A) to the connected feeders (B). 2.4.4. HMBPP dose-response evaluation Different dosages of HMBPP (1 µM and 100 µM) were evaluated in addition to the standard 10 µM that was used throughout the study, in order to determine if the Fig.2.5.b. Microscopic image of a fully fed female anopheline mosquito A B Fig.2.5.a. Microscopic image of an unfed female anopheline mosquito. - 29 - HMBPP-induced feeding rate was dose dependant. Feeding rate experiments were conducted as per the previously stated SMFA protocol (Section 2.4.3), using the two major vectors, An. funestus s.s. and An. gambiae s.s. Each biological replicate consisted of three technical replicates/HMBPP concentration. A total of three biological replicates was completed. The feeding rate was established by recording the number of fed females per concentration of HMBPP evaluated. i. Data analysis A dose-response graph was constructed using GrapPad Prism (9.0.1) and statistical difference between concentrations, for each species, was determined with the paired t-test (Kim, 2015) as both sets of data was normally distributed as confirmed with the Shapiro-Wilk test (Hanusz et al. 2016). In order to determine the statistical difference between the two species, at each concentration, the unpaired t-test was used. 2.5. Artificial infection of different mosquito species with Plasmodium falciparum 2.5.1. Infection through SMFA In order to determine if HMBPP has an effect on the susceptibility of major (including An. arabiensis, An. gambiae s.s., An. coluzzii and An. funestus s.s), minor (An. merus) and non-vector (An. quadriannulatus) mosquitoes to P. falciparum, female mosquitoes were fed on human blood containing mature gametocyte culture. A gametocyte culture consisted of Stage V gametocytes from the NF54 parasite strain. The gametocytemia of cultures varied between 1.5-2.5%, with a 50% hematocrit in A+ male serum with added fresh red blood cells (blood group was dependent on the donor) (Reader et al. 2021). Prior to setting up the SMFA, the gametocyte culture was assessed to ensure exflagellation of male gametes and the presence of a 3:1 female:male ratio. The gametocyte culture was provided by the Biochemistry Department of the University of Pretoria. The same protocol was followed as per section 2.4.3. to conduct feeding rate assays through SMFA. Each experiment consisted of three technical and biological replicates of which the control group was - 30 - supplied with human blood, containing gametocyte culture only, whereas the test group received gametocyte infected blood with added HMBPP (10 µM/1 ml blood) (Merck (Sigma-Aldrich); CAS number 396726-03-7; South Africa). Each cup consisted of 25 An. gambiae complex (An. gambiae s.s.; An. coluzzii; An. arabiensis; An. merus and An. quadriannulatus) and 30 An. funestus group (An. funestus s.s.) females (5-7 days old) (Fig.2.6). Female mosquitoes accessed the blood by piercing through the membrane and were allowed to feed for 40 minutes in dark conditions. Mosquitoes were starved for 2 – 3 hours prior to feeding. All unfed mosquitoes were removed with a mouth aspirator and fully fed mosquitoes were housed in a BSL2 facility of the Maureen Coetzee insectary, at a temperature of 25C and 80% relative humidity and a 12 hour day/night cycle with a 45 dusk/dawn transition of 45 minutes. Mosquitoes were provided with a 10% sugar water solution, containing 0.05% 4- aminobenzoic acid (PABA) (Merck (Sigma-Aldrich); CAS number 150-13-0; South Africa), until eight days post-infection. Fig. 2.6. Mosquitoes gathering at the feeder and piercing through the membrane, feeding on HMBPP-supplemented gametocyte culture. 350 ml mosquito feeding cup 1 ml glass feeder Circulating water - 31 - 2.5.2 Midgut dissections On days 8 – 10, post-infection, mosquitoes were immobilized and with the use of dissecting needles, the midguts were removed from infected specimens. Mosquitoes were dissected in PBS and guts were subsequently transferred to a droplet of 0.1% mercurochrome (Merck (Sigma-Aldrich); CAS number 129-16-8; South Africa). For better visibility, guts were stained between 8 to 10 minutes prior to visualisation under the microscope. After staining, the presence of oocysts was easily detected, as oocysts stained a darker shade of red, in comparison to the rest of the midgut.The prevalence (presence of oocysts) (Fig. 2.7a) and intensity (number of oocysts) were determined by viewing the guts under bright field illumination at 10x - 20x magnification (Delves & Sinden, 2010). Mosquitoes that had no oocysts in the midgut therefore had a prevalence of 0 (Fig. 2.7b) Fig. 2.7.a. Midgut of a major vector, An. funestus, containing a P. falciparum oocysts. P. falciparum oocyst - 32 - Fig. 2.7.b. Midgut of an uninfected An. funestus mosquito with no parasite oocysts. 2.5.3 Oocyst prevalence and intensity data capture and analysis Oocyst prevalence and intensity data were entered into Excel spreadsheets and the oocyst prevalence and intensity was determined for the control and treatment group according to each species. Data analysis was conducted in R-Studio (1.3.1093) and the Generalised Mixed Linear model was used to analyse oocyst prevalence and intensity data with high variance (Crawley, 2013). This model includes the semi- parametric model to evaluate data that is not normally distributed and therefore the model is not restricted to normally distributed data. Therefore, normality tests were not conducted for these assays. The multiple regression model was used to account for replication as a random factor. Results are reported as beta-estimate-, Chi- square- and P-values, with df=1. These values were entered into GraphPad Prism (9.0.1) in order to construct graphs displayed. - 33 - CHAPTER THREE- Results 3.1 Mosquito vector dynamics 3.1.1. Species identification and seasonal variation A total of 182 anopheline mosquitoes that were morphologically identified as belonging to the An. funestus group, showed amplification during the An. funestus multiplex PCR assay. Molecular species identification revealed that four members of the An. funestus group were present, An. parensis, An. leesoni, An. vaneedeni and An. rivulorum (Fig. 3.1). Twenty seven samples (15%) failed to amplify even after repeating the PCR assay and were not included in the data. This is possibly due to DNA degradation of the specimens in the field or specimens were morphologically misidentified and failed to amplify. Fig. 3.1. Agarose gel for the PCR identification of members of the An. funestus group. Lanes 1 and 28 indicate molecular weight markers whereas lanes 2, 3 and 4 are controls for An. funestus and negative controls for DNA extraction and PCR respectively. Lanes 5, 7, 12, 14-16, 19 and 23 are An. leesoni (146 bp) whereas lanes 6, 17-18, 21 and 24-25 indicates samples identified as An. vaneedeni (587 bp). Lanes 8-11, 13, 20, 22 and 26-27 contained samples which did not amplify during the assay. Seasonal variation in the relative abundance of members of the An. funestus group is represented in figure 3.2 and Table 3.1. Anopheles parensis was found to be the most abundant species over the 8-month collection period with a total of 64 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 500 100 - 34 - mosquitoes identified, which comprises 35% of the total sample size. Anopheles vaneedeni was the second most abundant species with a total of 51 specimens (28%) and together with An. parensis, these species were present from August 2019- March 2020 (Fig. 3.2). Fig.3.2. Relative species composition (n) of members from the An. funestus group (the minor vectors, An. parensis & An. vaneedeni and non-vectors An. leesoni, & An. rivulorum) from the Mamfene region in KwaZulu-Natal as collected over a period of 8 months. Table 3.1: The absolute number of anopheline species from the An. funestus group collected over an 8-month period from sites in Mamfene, KwaZulu-Natal. The percentage mosquitoes collected per collection method is indicated. 0 2 4 6 8 10 12 14 16 18 20 AUG SEP OCT NOV DEC JAN FEB MAR Sp e ci e s co m p o si ti o n ( n ) Month An. parensis An. leesoni An. vaneedeni An. rivulorum Variable Number of An. funestus group mosquitoes collected Month Aug 42 Sep 20 Oct 21 Nov 23 Dec 20 Jan 9 Feb 10 Mar 8 Collection method Clay pot 60% Modified plastic bucket 12% Tyre 19% Other methods(CO2 traps, human landing catches etc.) 9% - 35 - From October to December the abundance for An. parensis remained stable but started to decline from January (Fig. 3.2). Anopheles vaneedeni numbers also remained constant, after the spike in August, but the abundance started to decline in December (Fig. 3.2). The latter species numbers started to increase again from February. Overall a positive correlation was found between An. parensis and An. vaneedeni (r= 0.77, P=0.024) (Appendix C). Anopheles leesoni comprised 12% of the total number of An. funestus group collections. Furthermore, An. leesoni was only present in collections made from August to December, with the highest abundance during September (Fig. 3.2). Anopheles rivulorum was only present in September to December collections and its abundance increased monthly (Fig. 3.2). This species contributed 9% of the total specimens that amplified during the PCR analysis. No correlation was found between other members of the An. funestus group (Appendix C). The highest number of mosquitoes were collected during August (n=42) whereas fewer An. funestus group mosquitoes were collected during January (n=9) to March (n=8) (Table 3.1). Most collections came from clay pots (60%) and modified plastic buckets was the individual collection method that collected the least mosquitoes (n=12) (Table 3.1). Seasonal variation of the relative abundance for members of the An. gambiae complex is illustrated in Figure 3.3, Table 3.2. Anopheles arabiensis dominated collections with a total of 3,550 (99.4%) mosquitoes collected over the 8-month period while the minor vector, An. merus, contributed towards 0.5% of the total number. The non-vector species, An. quadriannulatus, only accounted for the remaining 0.1% of collections (data curtesy of the SIT group). The number of An. arabiensis decreased from August to September but started to increase again in October and November (Fig 3.3). For December, the An. arabiensis number was lower, but a sharp increase was observed for January. Although there was a decline in the number of An. arabiensis mosquitoes in February, the numbers increased again in March (Fig. 3.3). - 36 - Fig. 3.3. Relative species composition of members from the An. gambiae complex (the major vector, An. arabiensis, the minor vector, An.merus and non-vectorAn. quadriannulatus) from the Mamfene region in KwaZulu-Natal as collected over a period of 8 months. Table 3.2: The absolute number of anopheline species from the An. gambiae complex collected over an 8-month period from sites in Mamfene, KwaZulu-Natal. The percentage of mosquitoes collected per trap is indicated. The minor vector, An. merus, was only present in collections from November, December and March. In comparison to the major vector, the peak time for An. merus collections was in March, whereas the highest number of An. arabiensis specimens was collected in January (Fig.3.3). The second highest number of An. merus mosquitoes collected, was during December, when An. arabiensis collections were low in comparison to November and January (Fig.3.3). A total of 50% of An. 1 10 100 1000 AUG SEP OCT NOV DEC JAN FEB MAR Sp e ci e s co m p o si ti o n Month An. arabiensis An. quadriannulatus An. merus Variable Number of An. gambiae complex mosquitoes collected Month Aug 199 Sep 66 Oct 269 Nov 468 Dec 112 Jan 1087 Feb 557 Mar 815 Collection method Clay pot 48% Modified plastic bucket 2,4% Tyre 0,07% Other methods (CO2 traps, human landing catches etc.) 49,5% - 37 - quadriannulatus collections was made during August, while An. arabiensis collections were also lower (Fig.3.3). Collections for An. quadriannulatus also coincided with the months during which the collections for An. arabiensis were lower, but no correlation was found between the non-vector and the major vector. Collections for the An. gambiae complex mosquitoes were the highest during January (n=1087), with the least number of mosquitoes collected during September (n=66) (Table 3.2). As with the An. funestus group collections, the clay pots were responsible for most of the collections (48%), per single collection method, of An. gambiae complex mosquitoes (Table 3.2). A comparison between species of the An. funestus group and An. gambiae complex is presented in Fig. 3.4. The overall numbers of the non-vector (An. leesoni, An. rivulorum and An. quadriannulatus) and minor vector (An. parensis, An. vaneedeni and An. merus) populations remain low in comparison to that of the major vector. No correlation was found between the major vector, An. arabiensis, and minor- (An. parensis: r=-0.21, P=0.60; An. vaneedeni: r=-0.50, P=0.20; An. merus: r=0.27, P=0.51) or non-vector (An. leesoni: r=-0.80, P=0.64; An. rivulorum: r=-0.46, P=0.24; An. quadriannulatus: r=-0.19, P=0.65) species (Table 3.3). Fig. 3.4. The relative abundance of the major vector, An. arabiensis and minor (An. parensis, An. vaneedeni, and An. merus) and non-vector (An. rivulorum, An. leesoni, and An. quadriannulatus) species from the An. funestus group and An. gambiae complex over an 8-month period. A logarithmic scale displays relative abundance data. Average monthly rainfall (mm) is also indicated for the period. 0 10 20 30 40 50 60 70 80 90 100 1 100 Aug Sep Oct Nov Dec Jan Feb Mar R ai n fa ll (m m ) R el at iv e ab u n d an ce ( n ) Month An. parensis An. leesoni An. vaneedeni An. rivulorum An. arabiensis An. quadriannulatus An. merus Total rainfall - 38 - Table 3.3: Pearson’s correlation between the relative abundance of the major vector, An. arabiensis and all other minor- (An. parensis, An. vaneedeni and An. merus) and non-vector (An. leesoni, An. rivulorum and An. quadriannulatus) species from the An. gambiae complex and An. funestus group. Results are displayed as r- and two-tailed P-values for each species (P>0.05 indicates no significant difference). An. parensis An. leesoni An. vaneedeni An. rivulorum An. quadriannulatus An. merus An. arabiensis r=-0.21 P=0.60 r=-0.80 P=0.64 r=-0.50 P=0.20 r=-0.46 P=0.24 r=-0.19 P=0.65 r=0.27 P=0.51 In terms of rainfall, An. arabiensis numbers were low during the drier month (September) and increased again as rainfall increased in October and November (Fig. 3.4). The relative abundance of An. arabiensis was, however, at its peak in January, while the rainfall was much lower. The same was observed for March when An. arabiensis numbers were high during a drier period. No correlation was found between An. arabiensis abundance and rainfall (r= -0.06, P=0.88) (Table 3.4) or any of the other species (Table 3.4). Table 3.4: Pearson’s correlation between rainfall and mosquitoes from the An. gambiae complex and An. funestus group. Results are displayed as r and two-tailed P-values for each species (P>0.05 indicates no significant difference). Rainfall An. parensis An. leesoni An. vaneedeni An. rivulorum An. arabiensis An. quadriannulatus An. merus Pearson’s r=-0.05 r=-0.40 r=-0.27 r=0.67 r=-0.06 r=0.13 r=0.23 Correlation P=0.89 P=0.33 P=0.51 P=0.06 P=0.88 P=0.72 P=0.58 3.1.2. Wild caught mosquito infection profiling The ELISA assay confirmed the absence of P. falciparum in the wild female An. funestus group mosquitoes collected, including An. parensis, An. vaneedeni, An. leesoni and An. rivulorum. None of the samples had an optical density value higher than that of the cut-off value and thus all of the samples were regarded as negative. Representation of the optical density values is displayed in Fig. 3.5 for nine of the samples analysed. - 39 - Fig. 3.5. Screening of mosquitoes for P. falciparum infection. Optical density values are measured at 405 nM. Well A1 represents the positive control, which consists of purified recombinant protein, and wells A2-9 contain the wild caught samples. Insectary reared unfed female mosquitoes were used as negative controls, which was added to the last 7 wells of the plate, H6-12. Cut off value was calculated to be 0,12. 3.2 . Mosquito Blood Feeding Behaviour, Attraction and Parasite Host Manipulation In order to further investigate mosquito feeding behaviour and infection in the laboratory, an artificial feeding system was developed. Rodhain conducted the first artificial feeding of insects, via a membrane, in 1912 (Rutledge et al. 1964). The feeding system has been redesigned and to date it is one of the most common systems used (Rutledge et al. 1964). During this study the standard membrane feeding assay (SMFA) was used to evaluate if changes in feeding and attraction behaviour occurred in uninfected African mosquito vector species.. 3.2.1. Mosquito attraction and HMBPP 3.2.1.1. Choice chamber evaluation It was vital to establish if the choice chamber was sensitive enough to allow adult female mosquitoes to select between different food sources. Mosquitoes’ attraction was therefore evaluated between sugar water vs blood. The proportion of An. arabiensis mosquitoes that showed attraction towards the blood source was significantly higher (P=0.018) than those that had selected the sugar water (Table 3.5). The same was observed for An. funestus (P=0.032) (Table 3.5). - 40 - Table 3.5. Statistical analysis of the choice chamber evaluation. Cow blood and sugar water were the two sources used for attraction evaluation. Results are presented for two major vector species, (n=3). Species HMBPP-supplemented blood (Treatment) Blood only (Control) A n . a ra b ie n s is Sample size 60 60 Beta-estimate 31.1 30 Chisq value, Df=1 55 P-value 0.018* A n . fu n e s tu s Sample size 60 60 Beta-estimate 30.6 29.3 Chisq value, Df=1 4.59 P-value 0.032* *Denotes significant difference 3.2.1.2. HMBPP and attraction of different mosquitoes It was evident that the choice chamber design was sufficient, and I evaluated the possibility that female mosquitoes were more attracted to HMBPP-supplemented blood. The major vector, An. arabiensis, was more attracted to HMBPP- supplemented blood (P=0.0022) (Fig. 3.6, Table 3.6) whereas the opposite was observed for for An. coluzzii (P=0.0039) and An. gambiae (P=0.0023) (Fig. 3.7 & 3.8, Table 3.6). Anopheles funestus showed a similar response to HMBPP-induced blood as An. arabiensis, with increased attraction towards it (P=0.011) (Fig. 3.9, Table 3.6). Both minor vectors, An. merus and wild caught An. vaneedeni, showed higher attraction towards blood containing the metabolite (P=0.011, Fig. 3.10, Table 3.6 and P<0.001, Fig. 3.11, Table 3.6). The strongest significance was observed for An. vaneedeni. The non-vector, An. quadriannulatus, was more attracted towards the blood in the control group (P=0.012, Fig. 3.12, Table 3.6). - 41 - Fig. 3.6. Attraction of the major vector, An. arabiensis, females towards the control and treatment (HMBPP-supplemented blood) group. Error bars, ± SE Asterisks indicates significant difference (*P<0.05, **P<0.01, ***P<0.001). Fig. 3.7. Attraction of the major vector, An. coluzzii, females towards the control and treatment (HMBPP-supplemented blood) group. Error bars, ± SE Asterisks denotes significant difference (*P<0.05, **P<0.01, ***P<0.001). Fig. 3.8. Attraction of the major vector, An. gambiae, females towards the control and treatment (HMBPP-supplemented blood) group. Error bars, ± SE are indicated, and asterisks show significant difference (*P<0.05, **P<0.01, ***P<0.001). Fig. 3.9. Attraction of the major vector, An. funestus, females towards the control and treatment (HMBPP-supplemented blood) group. Error bars, ± SE Asterisks indicates significant difference (*P<0.05, **P<0.01, ***P<0.001). - 42 - Fig. 3.10. Attraction of the minor vector, An. merus, females towards the control and treatment (HMBPP-supplemented blood) group. Error bars, ± SE, are indicated and asterisks show significant difference (*P<0.05, **P<0.01, ***P<0.001). Fig. 3.11. Attraction of the minor vector, An. vaneedeni, females towards the control and treatment (HMBPP-supplemented blood) group. Error bars, ± SE, Asterisks denote significant difference (*P<0.05, **P<0.01, ***P<0.001). Fig. 3.12. Attraction of the non-vector, An. quadriannulatus, females towards the control and treatment (HMBPP- supplemented