do not necessarily reflect the views of UKDiss.com.
- Abstract 3
- List of abbreviations 4
- Background 4
- Reproductive Lifecycle: Adapting to Different Environments 7
- Female Oviposition Behaviour 8
- Early Embryogenesis 10
- Endochorion Melanization 11
- The Discovery of Moderate ERD Prior to Serosal Cuticle Formation 11
- Embryo Metabolism and Water Storage 12
- Middle Embryogenesis 13
- Serosal Cuticle Formation 13
- The Timing of Serosal Cuticle Formation 14
- The Role of Chitin in the Serosal Cuticle 15
- Culex quinquefasciatus Energetic Trade-offs in Mid-embryogenesis 17
- Late Embryogenesis 18
- The Process of Tyrosine-Mediated Cuticle Tanning 18
- The Role of Melanin in Serosal Cuticle-Mediated ERD 19
- Vector Control: Applying our Knowledge 21
- Targets for Intervention and the Feasibility of Genetic Control 22
- Chemical Control 24
- Chitin Synthesis as a Novel Target 27
- Targeting Chitin Synthesis Inhibitors to Cryptic Oviposition Sites 28
- Ovicide Viability and Challenges for the Future 31
- Conclusions 32
- References 33
- Supplementary Information 39
The transfer of genes from one generation of mosquito vectors to another relies on females and the eggs they lay onto ephemeral aquatic habitats. Such environments present many dangers to developing eggs, including a high risk of desiccation. When first laid, eggs possess a permeable two-layered eggshell consisting of an endo- and exochorion and will die if removed from moisture. However, during mid-embryogenesis, an extraembryonic membrane known as the serosa wraps the developing embryo and secretes the serosal cuticle, considerably reducing permeability and allowing eggs to survive outside of water. Because different mosquito species tend to oviposit onto different aquatic habitats, the relative risk of desiccation their eggs face varies considerably. The selection pressure to develop drought-resistant phenotypes therefore also varies, leading to interspecific differences in egg resistance to desiccation (ERD). Aedes aegypti, Anopheles gambiae and Culex quinquefasciatus eggs, for example, can avoid desiccation in dry conditions for ≥72, 24 and 5 hours respectively. Many differences in the organisation, composition and morphology of their eggshells have traditionally been proposed to explain this, though other factors, such as metabolism, likely also play a role. Ultimately, we can apply our knowledge of these topics to novel vector control strategies. Such strategies could be implemented during dry periods where mosquitos rely on the accumulation of dormant, desiccation-resistant eggs in the environment for their population to persist. As urban mosquitos often spread their eggs between numerous cryptic habitats, they cannot always be treated by habitat removal or costly insecticides. Therefore, novel ovicidal control schemes involving the targeting of the chorion with affordable NaOCl solutions or directing particular insect growth regulators to cryptic habitats by auto-dissemination are promising options for the future.
- List of abbreviations
LCA: Last common ancestor
GlcNAc: N-Acetyl glucosamine
GORO: GOlden cuticle + ROse eyes
IGR: Insect growth regulator
CSI: Chitin synthesis inhibitors
TH: Tyrosine hydroxylase
DDC: Dopa decarboxylase
ebony: N-β-alanyldopamine synthase
tan: N-β-alanyldopamine hydrolase
DCE: Dopachrome conversion enzyme
DHICA: 5,6-dihyrdoxyindole-2-carboxylic acid
Mosquitoes (Diptera: Culicidae) are the primary vectors of several significant diseases in predominantly tropical and sub-tropical countries (WHO, 2014; Becker et al., 2010). Currently, three of the most relevant mosquito vectors are Aedes aegypti, Anopheles gambiae and Culex quinquefasciatus.Between them, they transmit arboviruses which cause yellow fever, dengue and Zika, parasitic protozoans such as the malarial parasite Plasmodium falciparum and parasitic helminths such as Wuchereria bancrofti, the major cause of lymphatic filariasis (Marcondes & Ximenes, 2016; Samy et al., 2016; WHO, 2016; Clemons et al., 2010).
Anautogenous female mosquitoes are defined by the fact that they require a provision of amino acids and lipids from a blood meal to begin developing eggs (Clements, 1992). Following oogenesis and fertilization, they oviposit transparent eggs with a smooth endochorion and compound exochorion, through which water can freely move (Valle et al., 1999). At between 17-35% of total embryogenesis (depending on the species) cells of an extraembryonic membrane known as the serosa entirely envelopes the embryo and secretes the serosal cuticle (Vargas et al., 2014). The creation of this new three-layered eggshell occurs concomitantly with a huge reduction in water flux and the onset of the ability of eggs to survive outside of water (Farnesi et al., 2017). In any insect species, egg resistance to desiccation (ERD) can be achieved in three ways: (1) reducing the rate of water loss; (2) increasing the minimum water content that can be tolerated before death; and (3) increasing the whole water contents of the egg (Farnesi et al., 2015). Current literature suggests that modifications of eggshell components which relate to preventing water flux such as thickness, endochorion melanin content and serosal cuticle chitin content are the most relevant adaptions to ERD provision, though their relative contributions still remain to be discussed (Farnesi et al., 2017; Farnesi et al., 2015). It is also important to establish the supplementary roles that non-eggshell related factors, such as embryo metabolism, might play in desiccation resistance by increasing water storage (Gray & Bradley, 2005).
The degree of ERD at late embryogenesis varies between mosquito species, with Ae. Aegypti eggs being able to persist in dry conditions for ≥72 hours (high ERD) while An. gambiae and Cx. quinquefasciatus individuals in the same settings can only survive for 24 (medium ERD) and 5 (low ERD) hours respectively (Figure 1) (Farnesi et al., 2017). Interestingly, despite Ae. aegypti and Cx. quinquefasciatus being relatively more divergent in their levels of ERD, they are closely related in evolutionary terms (Ae. aegypti: Cx. quinquefasciatus LCA = 52-54mya; Ae. aegypti: An. gambiae LCA = 145-200mya). This demonstrates how the levels of desiccation resistance described in these species are the result of divergent ecological selection pressures imposed by their preferred oviposition sites, and not simply general evolutionary trends. Therefore, in order to discuss these species’ environment-specific adaptions, it is important to first understand the details of their ovipositional preferences and the unique environmental challenges which lie within.
Figure 1 Egg resistance to desiccation of Ae. aegypti, An. gambiae and Cx. quinquefasciatus on dry surfaces at late embryogenesis. Eggs at 80% of total embryogenesis were removed from water and transferred to 20-55% humidity dry conditions. Viability was then measured at regular intervals. ‘+’ indicates that Ae. aegypti have been known to survive desiccation in a dormant state for several months outside of water in some cases. Source of viability data: (Vargas et al., 2014). Adapted from Farnesi et al. (2017).
Cx. quinquefasciatus females exclusively oviposit onto permanent bodies of water which contain high levels of nutrient waste and microbial activity, such as wastewater ponds or flooded lavatories (Barbosa et al., 2007; Suleman & Shirin, 1981). An. gambiae also tend to oviposit onto permanent bodies of water. However, unlike Cx. quinquefasciatus, they prefer oxygen rich, clean water habitats such as rice fields, drainage canals or even tire prints (Sumba et al., 2004; McCrae, 1984). In contrast, Ae. Aegypti are described as ‘floodwater’ mosquitoes as they lay their eggs onto moist soil, tree holes or containers which periodically catch rainfall, and often must survive in low-humidity conditions for long stretches in a dormant state before flooding allows them to hatch (Reiter, 2007; Sota & Mogi, 1992). Interestingly, during the dry season in sub-Saharan Africa, Anopheles species are also often forced to oviposit onto moist floodland soil in the absence of permanent water bodies. While their eggs can only survive on dry surfaces for one day, these moist soils allow them to maintain viability for up to eighteen days in anticipation of flooding (Jawara et al., 2008; Shililu et al., 2004; Beier et al., 1990). Ultimately, while desiccation risk appears to be the primary factor driving differences in the physiology of species’ eggs, it is important to identify other environmental factors, such as the presence of pathogens or hypoxic conditions, may be forcing species such as Cx. quinquefasciatus to trade-off energetic investment from desiccation resistance into other processes such as gas exchange or immunity.
In the past, the only highly effective control schemes for diseases, such as malaria and dengue, have focused on reducing the population of vectors themselves through sterile insect technique, habitat removal, supplying bednets or insecticide spraying regimes (Raghavendra et al., 2011; Gerstl et al., 2010; Beier, 2008; Barat, 2006; Ooi, Goh & Gubler, 2006; Ault, 1994). Unfortunately, in spite of the fact that the developing egg represent a highly vulnerable stage of the mosquito lifecycle, embryogenesis remains its least studied developmental stage and very few examples of applied ovicidal vector control exist (Beament, 1989). By understanding the complexities of ERD, we can consider new vector control methods which seek to interrupt vital structures and processes, generating mosquito eggs that can no longer resist the pressures of their environment and die. Ultimately, this report aims to explore the possibilities for egg-targeted control, evaluating the current viability of a genetics-based approach and suggesting novel ovicidal chemicals such as NaOCl solutions and chitin synthesis inhibitors, which exploit our knowledge of the structures/ processes involved in desiccation resistance.
- Reproductive lifecycle: Adapting to different environments
Despite every mosquito species having a reproductive lifecycle that is unique to its own ovipositional niche, it is possible to describe a generalised process of oogenesis and embryogenesis for the Culicidae family as a whole. In all mosquito species, oogenesis involves a three step process of (1) pre-vitellogenesis, a pre-bloodmeal resting stage; (2) vitellogenesis, wherein nutrients are deposited into the oocyte to form the yolk; and (3) choriogenesis, wherein the endo- and exochorion are produced (Papantonis, Swevers & Iatrou, 2015; Swevers et al., 2005). The structure of the developing ovarian follicle during oogenesis is a self-contained unit consisting of the oocyte, 7 nurse cells and a surrounding follicular epithelium (Clements, 1992). Eggs are ultimately fertilised during the process of oviposition itself, meaning that the conserved stages of early, middle and late embryogenesis all occur outside of the mother and in the presence of environmental stressors (Degner & Harrington, 2016). At each stage of egg development, species accumulate differences which protect against the unique pressures of their habitats. These differences are best explored by analysing mosquito development chronologically (Fig.4).
4.1 Female oviposition behaviour
Mature female mosquitoes are considered gravid (egg bearing) when their oocytes are replete and possess a chorion. At this point, they begin oviposition searching behaviour in concordance with their daily activity patterns (Day, 2016). Typically, mosquitoes have fairly rigid patterns of circadian activity and only tend to oviposit at dawn or dusk. However, environmental factors such as rain, temperature and relative humidity can significantly affect where they fly and ultimately oviposit (Bidlingmayer, 1974). It can take several months for a gravid female to find a suitable oviposition site as they use a variety of long- (meters), mid- (meters to centimetres) and short-range (centimetres to contact) cues to evaluate potential sites (Bentley & Day, 1989). This allows mosquitoes to be exacting in their site evaluations, an ability which is essential as the quality and suitability of egg environments ultimately determines reproductive success.
At long range, visual cues help mosquitoes locate bodies of water against terrestrial backgrounds. The specifics of this behaviour are not fully understood, though it has been suggested that both ultraviolet and polarised light play a role in orientation towards potential sites (Silberglied, 1979; Wellington, 1974). Mosquitoes which oviposit at night (Cx. quinquefasciatus and An. gambiae) have also been shown to visually identify water sources by the infrared radiation released at dusk following absorption of heat during the day (Gibson, 1995). In lab experiments, Ae. aegypti, An. gambiae and Cx. quinquefasciatus all favour darker oviposition waters (Zuharah, Thiagaletchumi & Fadzly, 2016; Panigrahi et al., 2014; Huang et al., 2007). In Ae. aegypti especially, this preference is understood to be a reflection of their natural oviposition sites since soil becomes darker when wet and light cannot enter below or to the side of containers or tree holes, making them appear dark from above (Williams, DeLong & Venard, 1961).
As females approach potential oviposition sites, olfaction allows for more precise distinctions between site-specific blends of olfactory signals (Du & Millar, 1999). Anopheles, for example, are attracted to volatiles emanating from wild grasses which often occupy their preferred habitats (Asmare et al., 2017). Similarly, Cx. quinquefasciatus are significantly attracted by the olfactory signals emitted from a variety of decaying organic matter (Allan, Bernier & Kline, 2005). Upon contact, mosquitoes are hypothesised to evaluate site quality through the use of chemotactile receptors (Day, 2016). For example, when ovipositing onto moist soil, Ae. aegypti and An. gambiae females utilise moisture-sensitive receptors to evaluate water content, preferring soil patches with ≥70% soil saturation moisture content (Knight & Baker, 1962).
In regards to oviposition itself, Ae. aegypti lay their eggs individually around the edges of oviposition sites, just above the high water mark. An. gambiae also deposit their eggs individually, though they tend to scatter them while hovering (Day, 2016). Their eggs also notably possess lateral floats which allow them to sit on top of the water (Lounibos et al., 1997). Comparatively, Cx. quinquefasciatus lay their eggs in organised rafts of between 100-300 individuals. This structure grants protection from desiccation by minimising individual exposure to the external environment (Clements, 1992; Christophers, 1945). Throughout the entire Culicidae family, eggs can be classified into one of two classes: rapid hatch and delayed hatch. Rapid hatch eggs, which are laid directly onto water, see larvae eclose (hatch) immediately after embryogenesis concludes (e.g. Cx. quinquefasciatus). Comparatively, delay hatch eggs are laid onto moist surroundings and eclosion occurs only after flooding (e.g. Ae. aegypti) (Day, 2016). Interestingly, An. gambiae females tend to lay a mixture of both classes, adapting to lay a higher ratio of delay hatch eggs in dry months as a scarcity of suitable aquatic habitats forces them to oviposit onto moist floodland soil (Beier et al., 1990). Both An. gambiae and Ae. aegypti also practice ‘skip oviposition’ as they scatter their eggs between multiple sites (Okal et al., 2015; Colton, Chadee & Severson, 2003). This activity is likely a precautionary measure which ensures that females do not invest too much reproductive potential into any one site when dry conditions may result in it becoming unviable after egg laying. In the case of An. gambiae, females only tend to practice skip oviposition in around a third of cases as they can rely on permanent bodies of water for most of the year (Okal et al., 2015). I suggest that a future experimental target should be the determination if, during dry periods, An. gambiae practice skip oviposition at a higher frequency as this would have implications on the use of chemical control strategies which focus on auto-dissemination (See section 5.2.2.).
4.2 Early Embryogenesis
Following copulation, anywhere between 30-300 mature eggs can be laid by a female at one time (Rozendaal, 1997). Immediately after oviposition, mosquito eggs are transparent and have an eggshell consisting of two key layers, a smooth endochorion and a compound exochorion, both of which are synthesised by follicle cells during choriogenesis (Rezende et al., 2008). Because water can move through this chorionic bilayer, eggs initially increase in weight and overall size (Rezende et al., 2008; Kliewer, 1961). A so-called “inner eggshell” consisting of a homogenous vitelline membrane and a wax layer has also been described in many insects (Rezende et al., 2016). However, since mosquitoes have been shown to lack this wax layer and their vitelline membrane primarily functions in storing pro-proteins for the chorion, studies that investigate desiccation resistance mostly omit this layer from their discussions (Marinotti et al., 2014; Monnerat et al., 1999). Regardless, it has been included in Fig.4 to accurately represent the complexity of eggshell layers.
4.2.1. Endochorion melanization
The first modification of early embryogenesis involves the darkening of the endochorion by melanization (Clements, 1992). The enzymes involved in this process are deposited (as readily-active enzymes or zymogens) into the endochorion by follicle cells during choriogenesis and act to melanize and/ or sclerotize the eggshell when eggs are laid (Rezende et al., 2016). Here, sclerotization defines the formation of rigid sclerotin via the cross-linking of particular cuticle proteins (CPs) (Sugumaran, 1987). Ultimately, the degree to which different species’ eggshells are melanized can be seen in their colouration with Ae. aegypti, An. gambiae and Cx. quinquefasciatus eggs possessing black, dark brown and lightly tanned eggshells respectively (Fig.4) (Farnesi et al., 2017). Interestingly, by late embryogenesis, these degrees of eggshell melanization correlate to serosal cuticle-dependent ERD as darker eggs resist more to desiccation (Farnesi et al., 2015). The ways in which this phenomenon might be conferred are discussed in section 4.4.2.
4.2.2. The discovery of moderate ERD prior to serosal cuticle formation
Mosquito eggs were previously thought to be unable to resist desiccation prior to the formation of the serosal cuticle. However, more recent studies have challenged the simplicity of this view. For example, if removed from moisture immediately after oviposition, Cx. quinquefasciatus eggs can maintain ̴20% viability after 2 hours of exposure to dry conditions and Ae. aegypti eggs are able to maintain 30-50% viability after a considerably longer period of 10 hours (Farnesi et al., 2017). Interestingly, despite possessing an intermediate level of ERD compared to these two species at late embryogenesis, An. gambiae resistance to desiccation is null prior to serosal cuticle formation (Farnesi et al., 2017). Therefore, Anopheles’ dark eggshell may instead be related to camouflaging eggs against their dark backgrounds or providing protection from deleterious UV rays from the sun (Abram et al., 2015; Hinton, 1981). The question we must now ask is, in the absence of the serosal cuticle, what other factors could be granting certain eggs this moderate resistance?
Eliminating some possibilities, the exochorion is known to not play a role in ERD as it has been thoroughly established that its removal by treatment with NaOCl solutions have no effect on the ability of eggs to survive desiccation at any stage (Farnesi et al., 2015; Jacobs et al., 2013; Christophers, 1960). This layer instead appears to be important for shaping the eggs as well as providing physical protection in a way that permits maximum gas exchange (hence its lamellar structure) (Crampton, Beard & Louis, 2012). Moreover, effects of endochorion melanization can be ruled out because of the previously described observation wherein An. gambiae eggs possess a substantially darker eggshell than Cx. quinquefasciatus, yet have no ability to resist desiccation until serosal cuticle formation (Farnesi et al., 2017).
4.2.3. Embryo Metabolism and Water Storage
One largely unexplored explanation for differences in ERD is the relative availabilities of different metabolites such as glycogen, fatty acids and trehalose inside the egg (Farnesi et al., 2015). Adult Anopheles arabiensis, for example, have been shown topractice adapted levels of glycogen-trehalose conversion which lead to higher glycogen and free fatty acid content and lower trehalose content compared to its less desiccation resistant sister species, An. gambiae (Gray & Bradley, 2005; Wyatt, 1967). Since both water content at death and overall rates of water loss were the same between both species, these differences in metabolism were suggested to be the primary driving factor behind differences in ERD (Gray & Bradley, 2005). Here, glycogen would confer ERD by binding water up to five-times its own weight and storing it intracellularly, resulting in mosquitos which possess a higher percentage of water content (Schmidt-Nielsen, 1997; Gibbs, Chippindale & Rose, 1997). Adding to this, comparative studies involving Ae. aegypti, Aedes albopictus, and Aedes paullusi adults also found that high amounts of both accumulated glycogen and lipids were correlated with longer survival in dry conditions (Sawabe & Mogi, 1999). Ultimately, components of metabolism seem to be necessary in preventing desiccation prior to serosal cuticle formation in some species and also go some way to explaining the subtle differences between closely related species with no discernible eggshell differences. The role of metabolism in ERD should now be confirmed by comparing the availabilities of metabolites such as glycogen in the mosquito eggs of species which are known to possess moderate levels of pre-serosal cuticle ERD (e.g. Ae. aegypti) to those who are not (e.g. An. gambiae).
4.3 Middle Embryogenesis
The formation of the serosal cuticle during mid embryogenesis has been shown to occur concomitantly to a considerable reduction in water flux, granting mosquito eggs the ability to survive desiccation outside of water (Farnesi et al., 2017; Vargas et al., 2014). Variation in this cuticle’s physiology appears to be key to the divergent levels of desiccation resistance between species.
4.3.1. Serosal cuticle formation
The serosa is an extraembryonic membrane which is first defined during the differentiated blastoderm stage of early embryogenesis (Panfilio, 2008). During mid embryogenesis, serosal cells completely wrap the developing embryo and begin secreting the serosal cuticle as the germband first extends then retracts, initially defining the embryo’s body segments (Fig.4) (Goltsev et al., 2009; Handel et al., 2000). Studies in the grasshopper Melanoplus differentialis describe this cuticle as consisting of two layers, an endocuticle containing chitin and a lipoid epicuticle impregnated with a wax-like substance (Rezende et al., 2016; Slifer, 1937). This two-layered structure has not yet been demonstrated specifically in mosquitoes, though preliminary findings do indicate its existence (Rezende et al., 2016). Unfortunately, it is difficult to evaluate many of the complexities of the mature serosa, such as its thickness relative to other layers, because it is not possible to remove either the serosal cuticle or the serosa membrane without affecting the overlying chorion eggshell (Jacobs et al., 2013). Nevertheless, many interspecific differences in the attributes of the serosal cuticle have been described.
4.3.2. The timing of serosal cuticle formation
The first observable difference between species is the relative time of serosal cuticle secretion (Vargas et al., 2014). It is not known with certainty what physically triggers this process, though peak expression of ecdysone (an active form of ecdysteroid) has been shown to occur at the onset of serosal cuticle formation, likely working autocrinally to trigger secretion by the serosal cells (Lagueux et al., 1979). When considered relative to total embryogenesis, serosal cuticle formation occurs between 18.1-20.7%, 17.5-21.4% and 29.1-35% respectively in Ae. aegypti, An. gambiae, Cx.
quinquefasciatus (Fig.2) (Vargas et al., 2014). Hence, there appears to be a correlation between ERD and the speed of serosal cuticle formation.
Figure 2: Periods of egg darkening and serosal cuticle formation relative to total embryogenesis time. Colour of bars are representative of the complexion of each species’ eggshell following darkening. Exact times of total embryogenesis are as follows: Ae. aegypti = 77.4hrs; An. gambiae = 51.3hrs; Cx. quinquefasciatus = 34.2hrs. Exact timing of serosal cuticle formation after egg laying are as follows: Ae. aegypti 14-16hrs; An. gambiae = 9-11hrs; and Cx. quinquefasciatus = 10-12. Estimates for darkening time from (Clements, 1992; Christophers, 1960). Serosal cuticle formation information from (Vargas et al., 2014). Based on source: (Farnesi et al., 2017).
4.3.3. The role of chitin in the serosal cuticle
A variety of fluorescent labelling techniques have now demonstrated that the serosal endocuticle of mosquito eggshells is chitinised (Rezende et al., 2008). Chitin itself is a polysaccharide formed of β-1, 4 linked units of N-Acetyl glucosamine (GlcNAc) (Dutta, Dutta & Tripathi, 2004). Other than eggshells, it is also found in the adult mosquito exoskeleton and peritrophic matrix of the midgut (Merzendorfer & Zimoch, 2003; Clements, 1992). Interspecific differences in serosal cuticle chitin content have also been defined with the degree of chitinisation (GlcNAc (ng) / 1g eggshell) in Ae. aegypti (0.674), An. gambiae (0.317) and Cx. quinquefasciatus (0.182) correlating to their relative levels of ERD (Farnesi et al., 2015). The presence of chitin may confer ERD because, by virtue of being a long-chain polysaccharide, the microfibers it forms are hydrophobic to water (Dutta, Dutta & Tripathi, 2004). However, as will be explored in section 4.4.2., chitin may also produce its effect by taking part in cross-linking reactions at the endochorion-serosal cuticle border.
All insect species possess two classes of chitin synthase genes, CHS-A and CHS-B, which are thought to originate from a duplication event in a common ancestor(Merzendorfer & Zimoch, 2003). CHS-A genes are typically able to produce two splice variants which are expressed at different life-cycle stages (Merzendorfer, 2006). In Aedes species, the chitin synthase genes AaCHSIa/b (Class A) and AaCHSII (Class B) have both been shown, through transcriptomics, to be expressed in the mature serosa and have even been visualised by in-situ hybridisation in isolated eggshells (Table 2) (Jacobs et al., 2013). Studies have also found that, at the exact moment of SC formation, AaCHSIa is the sole variant expressed, suggesting that the chitin found in the serosal endocuticle is synthesised by this variant alone (Rezende et al., 2008). In the beetle Tribolium castaneum, injection of RNAi for the specific chitin synthase gene variant TcCHS1a produces eggs that have disorganised, de-chitinised serosal cuticles and are significantly less viable in low-humidity environments (Jacobs et al., 2013). Similar effects are conferred with knockouts of the genes Knickkopf1 or Radioactive, which are necessary for proper chitin microfibril formation (Jacobs et al., 2015).
4.3.4. Culex quinquefasciatus energetic trade-offs in mid-embryogenesis
The existence of a serosal cuticle is an ancestral trait in all mosquitoes (Jacobs et al., 2013). However, in the case of Cx. quinquefasciatus, a penchant for large permanent pools of water has likely resulted in their eggs experiencing negligible selection pressure to develop a waterproofing eggshell over the course of their evolution (Clements, 1992). Rather, the pressures imposed by Cx. quinquefasciatus’ polluted egg-stage habitats relate to low oxygen availability and high levels of pathogens (Barbosa et al., 2007). It is for this reason that the species appears to have evolved to deviate energy away from processes involved in serosal cuticle formation in exchange for investment into other, more relevant, processes.
For example, beyond its role in forming the serosal cuticle, the serosa membrane has been described as an “immune tissue” that is capable of generating an immune response against external pathogens (Vargas et al., 2014; Jacobs et al., 2013). Since the danger of pathogens in the egg-stage habitats of Cx. quinquefasciatus likely imposes a greater selection pressure on their eggs than the risk of desiccation, an energetic trade-off biasing investment into immunity may be occurring.
Moreover, the scavenging of dissolved oxygen from anoxic environments is also likely more crucial to the survival of Cx. quinquefasciatus eggs than resisting desiccation (Lane & Crosskey, 2012). Since the presence of the serosal cuticle inherently impedes gas exchange, delaying its formation would allow for maximum time spent freely exchanging gas through the two-layered eggshell which, in turn, allows for relatively fast development. (Vargas et al., 2014; Woods, 2010). More generally, a weak serosal cuticle would also permit greater gas flux from mid-embryogenesis to hatching (Farnesi et al., 2015). Ultimately, I would hypothesise that the delaying of serosal cuticle formation as well as a weak final product are likely both manifestations of a trade-off between ERD and gas exchange/ immunity in Cx. quinquefasciatus. In comparison, high desiccation risk associated with the ovipositional habitats of Ae. aegypti and An. gambiae has led to their eggs evolving a highly impermeable serosal cuticle at the expense of immunity, gas exchange and developmental speed.
4.4 Late embryogenesis
It has been shown that the viability of An. gambiae and Cx. quinquefasciatus eggs outside of water continues to increase even after the serosal cuticle has been established, implying that the structure undergoes a degree of maturation before becoming fully functional. Recent studies now indicate that, despite conferring no effects at early embryogenesis, the degree of endochorion melanization is important in serosal cuticle-mediated ERD as darker eggshells correlate with higher ERD at late embryogenesis (Farnesi et al., 2017).
4.4.1. The process of tyrosine-mediated cuticle tanning
In order to explain this observation we must first understand the process of tyrosine-mediated cuticle tanning (melanization and sclerotization) in mosquitoes as a whole. Figure 3 contains an overview of this process and, for reference, a detailed overview can be found in the Supplementary Information (Arakane et al., 2016; Shamim et al., 2014).
4.4.2. The role of melanin in serosal cuticle-mediated ERD
The ‘melanization-ERD’ hypothesis, wherein greater eggshell melanization is predicted to confer higher levels of serosal ERD, has recently been confirmed through studies of the mutant GORO strain of Anopheles quadrimacultus (Farnesi et al., 2017). These mutants are unable to melanize effectively at any life cycle stage and so possess a light, golden colouration as opposed to the dark-brown wildtype (Mazur et al., 2001). Importantly, the timing of serosal cuticle formation as well as the overall length of embryogenesis in this strain are identical to the wildtype, making differences in melanization the single divergent factor between study groups in terms of egg development. Overall, viability studies found that, at 25% of embryogenesis completion (i.e. following serosal cuticle establishment), only 17% of GORO eggs hatched when left for five hours in a dry environment, compared with 85% of WT individuals (Farnesi et al., 2017). This demonstrates that eggshell melanization somehow grants the serosal cuticle its ability to promote egg survivorship outside water.
In terms of interspecific differences, when the exochorion is removed by bleach, the degree of melanization in the remaining eggshell (endochorion and serosal cuticle) is highest in Ae. aegypti and lowest in Cx. quinquefasciatus (Farnesi et al., 2017). It is not known whether this difference in melanization is purely owed to differences in the endochorion from early embryogenesis or if the serosal cuticle is also melanized as it matures. One suggestion in favour of this second hypothesis is that serosa cells have been shown to express tyrosine hydroxylase and dopa decarboxylase enzymes, which are involved in melanin biosynthesis. However, since these enzymes are also involved in known sclerotization reactions, this discovery is not conclusive (Fig.3) (Goltsev et al., 2009). Determining if the serosal cuticle is melanized would allow us to more specifically hypothesise about the molecular basis of the interactions between the endochorion and serosal cuticle which result in ERD. However, such a determination may prove challenging as physical removal of the chorion by bleach is likely to de-pigment the cuticle (Farnesi et al., 2017). Moreover, transcriptomic expression studies of the serosal cuticle would also be unviable because the relevant genes in this case have an early onset of expression, prior to the stage at which the serosa can be isolated by eggshell removal. Therefore, I suggest that in situ hybridisation tests akin to those which confirmed the expression of DDC and TH should be performed for other melanin biosynthesis enzymes which are theoretically exclusive to the melanization pathway, such as DCEs (Goltsev et al., 2009).
In the absence of this information, it is not known how exactly the process of eggshell melanization is physically able to impact water retention, though a number of theories warrant evaluation. One suggestion is that polymeric melanin is possibly hydrophobic (Rajpurohit et al., 2014). However, this position is based on many assumptions, primarily that products of hydrophobic monomers, such DHI and DHICA, would share their hydrophobicity. In actuality, when assembled into a polymer, melanin hydrophobicity is dependent on the specific, complex biochemistry behind its formation (the word “melanin” itself being a catch-all term for a variety of pigments) (Arakane et al., 2016). Therefore, whilst there are reports of polymeric melanin being water-insoluble (Shamim et al., 2014), there are also reports which demonstrate it to be soluble (Farnesi et al., 2017; Wigglesworth, 1957). Understanding this concept is important in contextualising previous reports which argued against the melanization-ERD hypothesis because selection for Drosophila with darker cuticles had no effect on desiccation resistance (Rajpurohit et al., 2016). As Farnesi et al. (2017) explain, this may simply be because ‘hydrophilic melanins’ were being selected for in this case.
A second suggestion is that endochorion melanin deposits interact with biomolecule components of the serosal cuticle (King & Sinclair, 2015; Parkash, Rajpurohit & Ramniwas, 2009; Kramer et al., 2001). Such phenomena would explain why mosquito eggs which have their exochorion and endochorion removed by bleach exhibit significantly more permeable serosal cuticles (Beckel, 1958). As mentioned before, knowing whether the serosal cuticle is melanized would impact our understanding of the relative involvement of each eggshell layer in these interactions. Regardless, even a transparent (i.e. unmelanized) serosal cuticle could still be closely associated with a melanized endochorion. Such interaction would occur by melanin deposits in the endochorion bonding covalently or non-covalently to proteins or chitin found in the serosal cuticle, crosslinking them to provide mechanical stiffness (Arakane et al., 2016; Clements, 1992). This action of melanin would resemble the polyphenolic networks in sclerotized adult insect cuticles (Riley, 1992; Sugumaran, 1987; Hackman & Goldberg, 1971). Melanin cross-linking has also already been implicated in granting impermeability to the cell walls of the fungi Colletotrichum lindemuthianum and in endowing mechanical hardness to the jaws of Glycera (Moses et al., 2006; Hutchison et al., 2002).
Overall, while it is a fact that less pigmented eggshells confer lower levels of ERD, it is still not known whether more lightly pigmented eggshells are also less sclerotized, nor do we know for certain the extent to which the serosal cuticle interacts with the endochorion via sclerotization, melanization, or both (Rezende 2017, Personal Communication). In either case, we currently cannot make definite conclusions about the nature of this endochorion-serosal cuticle interaction.
- Vector Control: Applying our knowledge
Applying what we know about mosquito egg development to vector control involves targeting the specific processes which permit egg survival in dry conditions, generating drought-intolerant mosquitoes which are unable to resist the pressures of their environment and die. Interrupting such functions could theoretically be achieved by genetic manipulation or the use of chemical ovicides. Based on what we know about the population dynamics of Ae. Aegypti and An. gambiae, they represent the most viable targets for control because they rely on accumulations of dormant, desiccation-resistant eggs in container/ floodland habitats (“egg banks”) to persist as a population through dry seasons (Jawara et al., 2008; Trpis, 1972). During dry periods, the populations of both these species are known to decrease dramatically as a result of a reduction in the quality and overall number of ovipositional sites available, leaving them to vulnerable to local extinctions if targeted (Mackay et al., 2015; Shililu et al., 2004). By pre-emptively targeting dormant egg stages, the rapid mosquito population increase at the beginning of wet seasons could be moderated (Baber et al., 2010). Moreover, these tactics could be used in “crash and release” strategies which improve the success of large-scale release programmes of genetically modified or Wolbachia-infected mosquitoes at the onset of wet seasons (Jacups et al., 2013). Ultimately, while the targeting of mosquito eggs during dry periods is one that has been suggested many times, practical suggestions are ultimately underexplored.
5.1. Targets for intervention and the feasibility of genetic control
Recent proteomic and transcriptomic analyses of Ae. aegypti and An. gambiae eggshells have provided an insight into their specific protein composition (Marinotti et al., 2014; Amenya et al., 2010). Tables 1 & 2 contain a summary of the genes identified by these studies which may be relevant to ERD, e.g. those involved in chorion production, melanization/ sclerotization and chitin synthesis. By analysing the expression of these genes and drawing from our knowledge of their roles in embryogenesis, we can consider their suitability for genetic control.
One promising development in the field of genetic control are CRISPR-Cas9 gene drive systems, which utilise site-specific endonucleases to spread female-recessive sterility in vector populations in order to suppress them. One target which was evaluated for use in this system is the Dopa-decarboxylase enzyme (DCE) AGAP005958 (Hammond et al., 2015). An ortholog of yellow-g in Drosophila, this gene playS a role in melanizing the endochorion during early embryogenesis by catalysing the conversion of Dopaminechrome to DHICA (Fig.4) (Suppl. Info) (Arakane et al., 2016; Hammond et al., 2015). Interestingly, the protein product of AGAP005958 was not found during proteomic analysis of the An. gambiae eggshell whilst its paralog (AGAP005959 (yellow-g2)) was (Table 1). However, as suggested by the authors, this may have been a result of methodological limitations (Amenya et al., 2010). Indeed, in Ae. aegypti, proteomics have confirmed the presence of both yellow-g and yellow-g2 orthologues (AAEL010848 & AAEL007096/AAEL002333) in eggshells (Table 2).
Interestingly, female AGAP005958 knockouts are unable to lay eggs (Hammond et al., 2015). While the physiological causes of this phenotype are currently unknown, analysing yellow-g’s role during embryogenesis may help us make informed suggestions. Firstly, one important fact to establish is that, despite mosquito eggs often being described in the literature as ‘transparent’ upon oviposition, a low level of melanization may occurs prior to oviposition, with darkening periods overlapping egg laying to conclude at early embryogenesis (Clements, 1992). During oogenesis in Drosophila, the yellow family of genes was suggested to be involved in “chorion formation and hardening” (Li & Christensen, 2011). Therefore, considering the tendency of mosquitoes eggs to utilise melanin-cross linking to reinforce structures during embryogenesis onwards, I suggest that one possibility is that low levels of melanin-mediated endochorionic protein cross-linking may also be occurring within the mother, granting unfertilised eggs the mechanical strength to survive until the point of hatching (even if such events have no effect on water retention after oviposition (Farnesi et al., 2017)).
Regardless, (Hammond et al., 2015) ultimately found AGAP005958 to be unviable as an option for gene drive because the negative effects its construct imposed on reproductive ability outweighed its honing rate. However, it was noted by the authors that this gene meets many requirements of female-specific RIDL (Labbé et al., 2012; Thomas et al., 2000). This system is an adaption of classic RIDL technique and involves breeding individuals to contain a tetracyclin-repressed transactivator (tTA) under the control of a suitable promoter as well as a dominant female lethal gene under the control of tetO (a tTA-response element) (Fig.5) (Labbé et al., 2012). During rearing, tetracyclin represses this system. However, when eggs are released into the field, the absence of tetracyclin causes females to die whilst RIDL-males emerge to spread this RIDL construct and confer the same effects in the following generation (Thomas et al., 2000). However, designing a system based around AGAP005958 would require development of a mechanism to allow for the conditional rescue of sterility, i.e. a gene/ construct which silences or regulates AGAP005958 would need to be identified, engineered and tested for its ability to allow AGAP005958 expression in the presence of tetracyclin, whilst repressing it naturally. Ultimately, it is hard to outline what advantages such a system would have over current schemes which have already been successfully shown to conditionally cause female flightlessness using the Actin-4 regulatory regions as a ‘lethal’ gene (Labbé et al., 2012). The same also applies to the suggestion of some authors that a novel RIDL scheme based on interfering with the components of ERD, especially since the relevant processes during embryogenesis do not seem to be exclusive to males or females.
5.2. Chemical Control
The physical removal of mosquito oviposition sites has been extensively practiced as a highly efficient means of vector control. However, in cases where this is not viable, sites can be treated with neurotoxic insecticides or insect growth inhibitors (IGRs) (Farnesi et al., 2012). Unfortunately, the use of neurotoxic insecticides is currently facing significant problems due to the ongoing development of population-level resistance (Liu, 2015; Hemingway & Ranson, 2000). As such, novel substances with alternative target sites are now needed (Zaim & Guillet, 2002).
As mentioned previously, pre-emptive ovicide-based vector control aimed at habitats that harbour delayed hatch eggs during dry seasons could have significant impacts on mosquito populations, inhibiting the increase in disease rates at the start of wet seasons (Mackay et al., 2015). The implementation of easy-to-use and affordable ovicides is especially required in cases where oviposition sites cannot be physically removed and/or larvicidal treatment is unviable due to sheer habitat abundance creating logistic and cost problems (Fernandez et al., 1998). It is also important to remember that, since dormant eggs make up a substantial proportion of many mosquito populations during dry seasons, ovicides would be much more effective than treatments which target larvae or even adults (Mackay et al., 2015). An example of one such ovicide is Sodium hypochlorite (NaOCl) solution, aka bleach, which can digest the chorion (Jacups et al., 2013). In field trails, 1:3 (13.125ppt) and 4:1 (42ppt) dilutions of household bleach mixed with a smectite clay thickening agent decreased the rate of egg hatching by over ≥98% in plastic containers and rubber tires/ concrete drums respectively (Mackay et al., 2015). When these containers were flooded, chlorine concentrations dropped below 2ppm, the concentration at which bleach is safe to drink and has no unpleasant taste (Mackay et al., 2015; WHO, 2011). The use of household bleach has therefore already been suggested as a community-led, cheap and easy-to-use method of killing the egg stages of mosquitoes (Mackay et al., 2015; Barrera, Amador & Clark, 2004). Unfortunately the use of such a scheme is limited in its scope and would not be viable as a large-scale control programme due to the sheer number of sites and the possible areas over which people would be required to spray.
5.2.1. Chitin Synthesis as a novel target
Chitin synthesis inhibitors (CSIs) are an example of an IGR class which could target ERD-granting processes specifically (Zhu et al., 2007; Cohen, 2001). In the eggs of Ae. aegypti and An. gambiae, chitin contributes to the water-proofing serosal cuticle by forming impermeable microfibers as well as likely taking part in interactions with melanins on the endochorion (Farnesi et al., 2015). It has been known since the 1970s that CSIs affect juvenile mosquito stages, though their specific modes of action are still not fully understood (Belinato et al., 2013). In larvae, they lead to the development of fragile cuticles, conferring individuals who are unable to shed their exuvaie and die by starvation or suffocation (Graf, 1993; Vasuki & Rajavel, 1992). Unfortunately, despite Drosophila studies demonstrating that CSIs such as lufernuron can also efficiently kill insect eggs, the practical targeting of the egg-stages of the mosquito lifecycle remains underexplored. (Mondal & Parween, 2000; Wilson & Cryan, 1997).
In Aedes and Culex mosquitoes, the CSIs diflubenzuron and triflumuron have been shown to act as larvicides (Martins et al., 2008; Fournet, Sannier & Monteny, 1993). Another promising candidate, novaluron, has also been shown to inhibit the growth of >99% of Ae. aegypti larvae in concentrations of just 0.5 – 1.0μgL-1, which is 6-12 times less than would be needed by triflumuron or diflubenzuron to have an effect (Farnesi et al., 2012). Encouragingly, many CSIs, including novaluron, have been approved for use in drinking water and, as such, their utilisation is already recommended by the WHO Pesticide Evaluation Scheme (WHOPES) in cases where insecticide resistance is present (WHO., 2007; Chavasse, Yap & WHO, 1997). This led to their implementation by the Brazilian national dengue control programme (Farnesi et al., 2012). I suggest that, as a future prospect, the relative effects of novaluron, diflubenzuron and triflumuron against the mosquito eggs of different species should be compared to evaluate their relative effectiveness. Experiments could also establish the effects of lowering humidity on egg survival at lower doses as partial disruption of the chorion may lead to more significant effects with regards to ERD.
5.2.2. Targeting Chitin Synthesis Inhibitors to cryptic oviposition sites
One possible application of our knowledge of both oviposition behaviour and the components of ERD is the targeting of skip-ovipositioning mosquito species with auto-dissemination stations containing CSIs. Such apparatuses contain water reservoirs treated with attractants to which gravid females congregate, along with a transfer chamber which contaminates these females with IGR-impregnated silica particles. Once contaminated, females transfer this IGR to secondary oviposition sites at lethal concentrations (Gaugler, Suman & Wang, 2012). Unlike neurotoxic insecticides, IGRs exclusively affect growing juveniles and so these gravid females suffer no deleterious effects that may impede disseminating behaviour. Ultimately, the use such “lure and disseminate” tactics have advantages over direct spraying or “lure and kill” strategies because their exploitation of the skip-ovipositing behaviour of mosquitoes such as Ae. aegypti, allows for targeting of “small volume cryptic habitats” which would otherwise prove difficult to target (Gaugler, Suman & Wang, 2012; Richards et al., 2006).
Putting these concepts into practice, a prototype station using the juvenile hormone mimic pyriproxyfen conferred 100% lethality of Ae. albopictus eggs within secondary sites in small cage trails (2.2 m3) and 81% inhibition in small room trials (31.1m3) (Gaugler, Suman & Wang, 2012). More recent field trials over areas as large as 0.8ha also found highly significant reductions (p<0.0001) in egg numbers as a result of treatment (Unlu et al., 2017).
When employed in the field, it is important to consider gravid female mosquito preferences in order to maximise efficiency (See: section 4.1). The stations designed by Gaugler et al., for example, mimicked local trees in their colour, texture and shape since Ae. albopictus mosquitoes preferentially oviposit onto tree-trunks (Fig.6). The station also contained attractants in the form of reservoirs with food substrates which facilitated bacterial growth and organic deposition (Gaugler, Suman & Wang, 2012). A goal for future designs should be the establishment of more species-specific attractants to reduce negative ecological effects because, despite gas-chromatographic studies helping to develop attraction profiles for Ae. aegypti and An. gambiae based on particular organic infusions, there are still no established compounds which would be specific to the level of species (Eneh, 2016; Ponnusamy et al., 2010). Overall, a single station can now be produced for <$1, making them a more cost-effective solution than manual habitat spraying (Gaugler, Suman & Wang, 2012).
Ultimately, whilst previous studies have focused on juvenile hormone-mimics as their chosen IGRs, CSIs possibly represent more effective alternatives. In fact, studies comparing different classes of IGRs found that low levels of resistance towards pyriproxyfen, but not novaluron, exists in in Ae. aegypti populations (Lau et al., 2015). This fact, combined with its high efficiency, led the authors to suggested that novaluron (alongside the neurotoxin Cyromazine) was the most promising IGR for the future. This could further be confirmed by comparative studies involving controlled novaluron and pyriproxyfen auto-dissemination cage and small-room trials.
5.3. Ovicide Viability and challenges for the future
Ovicide use in mosquito control has not been extensively practiced, and the reasons for this are likely multifaceted. The principal problem such applications face is that female mosquitoes have evolved to lay eggs far in excess of that which can be supported by ovipositional habitats. Therefore, because the majority of eggs are likely to die naturally, there is a buffer on the action of interventions. In contrast, during the larval and pupal stages, most individuals are likely to eventually develop to adulthood and so every individual killed will have more direct implications on the adult population. Despite this, there is still an argument to be made for the use of ovicides. Bleach spraying, for example, demonstrates a lethality rate which is so high that it overcomes this issue of buffering (>99%). Moreover, as mentioned before, cheap ovicidal interventions are sometimes the only effective option in cases where dedicated larvicide spraying is unviable or during dry months where mosquito population persistence is dependent on stocks of dormant eggs.
A major concern that is relevant to all chemical control schemes is a lack of species-specific activity leading to deleterious effects in non-target organisms. For example, chitin is known to be important in many organisms and so the introduction of chitin-targeting chemicals to an area could potentially harm a variety of biota (Tang et al., 2015; Arakane & Muthukrishnan, 2010). However, relative to neurotoxic insecticides, IGRs tend to show higher levels of safety with regards to non-target biota (Vythilingam et al., 2005). Field studies investigating novaluron treatment of larvae in Bangkok, for example, found no negative impacts on aquatic plants or fishes in treated areas up to three months after study conclusion (Tawatsin et al., 2007). Moreover, the intrinsic action of auto-dissemination in particular is one where stagnant, often small/ hidden sites are targeted and such habitats are unlikely to contain non-target organisms.
An issue which is relevant to any new vector control program is the possibility that, by eliminating a mosquito species by targeting their ovipositional niche, one simply opens up this niche for a new species to expand into. Our knowledge that closely-related mosquito species are often more divergent in their ovipositional niches than more distantly related species indicates that mosquito egg physiology is dependent on the aquatic sites available. This implies that any mosquito species which overlap in range with an eliminated population could eventually adapt and replace them, even if their ovipositional sites are currently divergent. This eventuality should not discourage the interventions recommended in this report however as the ultimate goal is not to eliminate populations but to act pre-emptily to dampen population recovery at the beginning of wet seasons
Overall, it is clear that the divergent selection pressures present at the preferred oviposition sites of different mosquito vectors has a large impact on many attributes of their eggs. Within this, the vital role of the chitinised serosal cuticle has been established in detail, though further research is needed to qualify how exactly its components interact with the melanized endochorion to confer mature levels of ERD. Regardless, it is becoming increasingly apparent that the traditional view of desiccation resistance as simply a product of the serosal cuticle is too reductionist as way of thinking. Rather, different species appear to resist desiccation through the action of many factors, into which they place different levels of importance. Such factors, particularly those which relate to the production of water-binding metabolites, explain why species such as Ae. aegypti and Cx. quinquefasciatus, possess a small degree of ERD prior to serosal cuticle formation. Finally, our knowledge of mosquito egg-stages suggest that current urban vector control regimes could be enhanced by introducing ovicidal treatments which target the standing crops of dormant eggs that allow mosquito populations to recover from population decreases in the dry season. Such treatments could involve such as community-led NaOCl container spraying programmes or even the supply of auto-dissemination stations which target cryptic oviposition sites with CSIs.
- Supplementary Information.
Biosynthesis of melanin and sclerotin:
The process of tyrosine-mediate cuticle tanning begins with the two-step transformation of phenylalanine → tyrosine → DOPA by hydroxylation reactions catalysed by phenylalanine hydroxylase (PO) and tyrosine hydroxylase (TH) respectively. The product, DOPA, can then be decarboxylated by the enzyme Dopa Decarboxylase (DDC) to form 3,4-dihydroxyphenethylamine (dopamine). In order to orchestrate sclerotization, dopamine is acetylated by N-acetyltransferase (NAT) or β-alanylated by N-β-alanyldopamine synthase (ebony) to yield N-β-alanyldopamine (NBAD) or N-acetyldopamine (NADA) respectively. These products are then oxidised by Laccase 2 to form NBAD-quinone or NADA-quinone which undergo cross-linking reactions with cuticle protein side chains for cuticle sclerotization. NBAD can also be converted back into dopmanie via N-β-alanyldopamine hydrolase (tan).
Both DOPA and dopamine can also be utilised to generate melanins. However, since DOPA is a relatively poor substrate for Laccase 2, the DOPA-melanin synthesis pathway often contributes little to melanization with the majority of polymeric melanin being synthesised via dopamine since, by comparison, its initial oxidation is highly efficacious. In either pathway, quinone products are cycled non-enzymatically to form dopachrome and dopaminechrome which act as substrates for Dopachrome conversion enzyme (DCE) to yield 5,6-dihyrdoxyindole-2-carboxylic acid (DHICA) and 5,6-dihydroxyindole (DHI). These then polymerise to DOPA-melanin/ dopamine-melanin which can be deposited (Arakane et al., 2016; Shamim et al., 2014).