Introduction to parasitic nematodes of mammals

Mammalian-parasitic nematodes infect over a billion people worldwide and are a major cause of morbidity in low-resource areas (Boatin et al., Reference Boatin, Basanez, Prichard, Awadzi, Barakat, Garcia, Gazzinelli, Grant, McCarthy, N'Goran, Osei-Atweneboana, Sripa, Yang and Lustigman2012). Infections with soil-transmitted nematodes can cause chronic gastrointestinal distress, stunted growth and cognitive impairment in children, anaemia and even fatality in infants and immunocompromised individuals (Lustigman et al., Reference Lustigman, Prichard, Gazzinelli, Grant, Boatin, McCarthy and Basanez2012). Infections with vector-transmitted nematodes can cause severe symptoms such as permanent disfigurement and blindness (Lustigman et al., Reference Lustigman, Prichard, Gazzinelli, Grant, Boatin, McCarthy and Basanez2012). Additionally, parasitic nematodes that infect livestock are an enormous source of economic burden (Jasmer et al., Reference Jasmer, Goverse and Smant2003). Current treatments for infections depend on anthelminthic drugs that reduce the worm burden in heavier infections but do not prevent reinfections, with the result that reinfection is common in endemic areas (Prichard et al., Reference Prichard, Basanez, Boatin, McCarthy, Garcia, Yang, Sripa and Lustigman2012). In addition, drug resistance resulting from mass drug administration is a major challenge for the treatment of nematode-infected livestock (Kumar et al., Reference Kumar, Rao, Varghese and Rathor2013; Roeber et al., Reference Roeber, Jex and Gasser2013; Emery et al., Reference Emery, Hunt and Le Jambre2016; Learmount et al., Reference Learmount, Stephens, Boughtflower, Barrecheguren and Rickell2016) and is expected to be a concern for the treatment of nematode-infected humans in the near future (Keiser and Utzinger, Reference Keiser and Utzinger2008; Diawara et al., Reference Diawara, Schwenkenbecher, Kaplan and Prichard2013; Repetto et al., Reference Repetto, Ruybal, Batalla, Lopez, Fridman, Sierra, Radisic, Bravo, Risso, Gonzalez Cappa and Alba Soto2018). The drugs currently available are also not sufficient to eliminate human infections in all cases, at least by following the administration schedules under practice (Repetto et al., Reference Repetto, Ruybal, Batalla, Lopez, Fridman, Sierra, Radisic, Bravo, Risso, Gonzalez Cappa and Alba Soto2018).

Many of these nematodes are gastrointestinal parasites with a developmentally arrested infective larval stage that inhabits the environment and infects hosts following either skin penetration or passive ingestion, depending on the species (Gang and Hallem, Reference Gang and Hallem2016; Bryant and Hallem, Reference Bryant and Hallem2018). The infective larval stages of these species respond robustly to a diverse array of host and environmental sensory cues, including CO₂ (Gang and Hallem, Reference Gang and Hallem2016; Bryant and Hallem, Reference Bryant and Hallem2018). In addition, many parasitic nematodes may rely on sensory cues inside the host body, including CO₂, to re-initiate development upon host entry, direct somatic migration and establish a successful infection (Hawdon and Schad, Reference Hawdon and Schad1990, Reference Hawdon and Schad1992; Hawdon et al., Reference Hawdon, Volk, Pritchard and Schad1992; Bekelaar et al., Reference Bekelaar, Waghorn, Tavendale, McKenzie and Leathwick2018, Reference Bekelaar, Waghorn, Tavendale, McKenzie and Leathwick2019).

Responses of skin-penetrating nematodes to CO₂

Skin-penetrating nematodes such as the human-parasitic hookworms Ancylostoma duodenale and Necator americanus and the human-parasitic threadworm Strongyloides stercoralis are gastrointestinal parasites that infect hosts as developmentally arrested third-stage larvae (iL3s) (Roberts et al., Reference Roberts, Janovy and Schmidt2005; Nutman, Reference Nutman2017; Velikkakam et al., Reference Velikkakam, Fiuza, Gaze and Singh2017). The iL3s are soil-dwelling and actively engage in host seeking using a variety of host-associated sensory cues (Gang and Hallem, Reference Gang and Hallem2016; Bryant and Hallem, Reference Bryant and Hallem2018). These parasites generally have narrow host ranges, infecting only a limited number of host species (Haley, Reference Haley1961; Bezubik, Reference Bezubik1965; Nolan et al., Reference Nolan, Zhu, Ketschek, Cole, Grant, Lok and Schad2007; Viney and Lok, Reference Viney and Lok2007; Viney and Kikuchi, Reference Viney and Kikuchi2017). After invading a host by skin penetration, the iL3s resume development inside the host, a process called activation (Stoltzfus et al., Reference Stoltzfus, Massey, Nolan, Griffith and Lok2012, Reference Stoltzfus, Bart and Lok2014). The nematodes then migrate through the host body to their final destination, the small intestine, where they reside as parasitic adults (Roberts et al., Reference Roberts, Janovy and Schmidt2005; Nutman, Reference Nutman2017; Velikkakam et al., Reference Velikkakam, Fiuza, Gaze and Singh2017). The adults reproduce in the small intestine, and then the eggs or young larvae, depending on the species, exit the host body in feces. The nematodes inhabit the feces until they develop into iL3s (Roberts et al., Reference Roberts, Janovy and Schmidt2005; Nutman, Reference Nutman2017; Velikkakam et al., Reference Velikkakam, Fiuza, Gaze and Singh2017). In most species, all of the progeny of the parasitic adults develop directly into iL3s. However, Strongyloides species can cycle through one or a limited number of free-living generations on the feces before developmentally arresting as iL3s (Roberts et al., Reference Roberts, Janovy and Schmidt2005). Strongyloides stercoralis can also pass through multiple generations inside the same host through autoinfective cycles (Roberts et al., Reference Roberts, Janovy and Schmidt2005).

Many skin-penetrating nematodes show behavioural responses to CO₂. For example, iL3s of the dog hookworm Ancylostoma caninum display increased nictation in the presence of CO₂ (Granzer and Haas, Reference Granzer and Haas1991). Nictation is a specialized behaviour displayed by many parasitic nematodes in which the worm stands on its tail and waves its head in the air to facilitate attachment to mobile hosts (Granzer and Haas, Reference Granzer and Haas1991; Bryant and Hallem, Reference Bryant and Hallem2018). In addition, both Ancylostoma caninum and Strongyloides stercoralis iL3s exhibit increased movement when exposed to human breath, and this behaviour is not observed when CO₂ is removed from the breath (Sciacca et al., Reference Sciacca, Forbes, Ashton, Lombardini, Gamble and Schad2002). Similarly, the human-parasitic hookworms Ancylostoma duodenale and Necator americanus display increased activity in response to CO₂ in combination with heat and/or moisture (Haas et al., Reference Haas, Haberl, Idris, Kallert, Kersten and Stiegeler2005). The similar responses of Strongyloides stercoralis and hookworms to CO₂ is particularly notable given their phylogenetic divergence, with Strongyloides stercoralis in clade IV and hookworms in clade V (Blaxter and Koutsovoulos, Reference Blaxter and Koutsovoulos2015; Blaxter et al., Reference Blaxter, Koutsovoulos, Jones, Kumar, Elsworth, Olson, Hughes and Cotton2016). However, these studies did not look at migration in CO₂ gradients, and whether CO₂ was an attractant or repellent was not clear.

More recent studies demonstrated that skin-penetrating iL3s of the human parasites Strongyloides stercoralis and Ancylostoma ceylanicum and the rat parasites Strongyloides ratti and Nippostrongylus brasiliensis are repelled by CO₂ in CO₂-chemotaxis assays (Fig. 1A, B) (Castelletto et al., Reference Castelletto, Gang, Okubo, Tselikova, Nolan, Platzer, Lok and Hallem2014; Ruiz et al., Reference Ruiz, Castelletto, Gang and Hallem2017). A lack of attraction towards CO₂ is consistent with the route of infection of skin-penetrating nematodes, since mammalian skin surfaces emit low concentrations of CO₂ (Alkalay et al., Reference Alkalay, Suetsugu, Constantine and Stein1971). On the other hand, fecal deposits contain high levels of CO₂ resulting from aerobic respiration of fecal bacteria (Jensen and Jorgensen, Reference Jensen and Jorgensen1994; de Lacy Costello et al., Reference de Lacy Costello, Amann, Al-Kateb, Flynn, Filipiak, Khalid, Osborne and Ratcliffe2014; Rotbart et al., Reference Rotbart, Yao, Ha, Chrispa, Muir, Gibson, Kalantar-Zadeh and Ou2017), and CO₂ repulsion may drive these iL3s off of host feces and into the environment in search of new hosts.

Fig. 1. Behavioural responses of mammalian-parasitic nematodes to CO₂. (A) A CO₂-chemotaxis assay. CO₂ and air are pumped into opposite sides of a 10 cm plate. Infective larvae (iL3s) are placed at the centre and allowed to migrate for 1 h. A chemotaxis index is then calculated according to the formula indicated; a positive index indicates attraction and a negative index indicates repulsion. (B) Behavioural responses of skin-penetrating iL3s to 10% CO₂. All species tested are repelled by CO₂. Data are from Castelletto et al. (Reference Castelletto, Gang, Okubo, Tselikova, Nolan, Platzer, Lok and Hallem2014) and Ruiz et al. (Reference Ruiz, Castelletto, Gang and Hallem2017). (C) Experience-dependent changes in CO₂ responsiveness in the passively ingested nematodes Heligmosomoides polygyrus and Haemonchus contortus. CO₂ responses switch from repulsion (in Heligmosomoides polygyrus) or neutral (in Haemonchus contortus) to attraction following removal from host feces for days to weeks. Heligmosomoides polygyrus was tested with 10% CO₂; Haemonchus contortus was tested with 15% CO₂. Figure adapted from Ruiz et al. (Reference Ruiz, Castelletto, Gang and Hallem2017). Graphs show medians and interquartile ranges. ****P < 0.0001, ***P < 0.001, Mann–Whitney test for each species.

Responses of passively ingested nematodes to CO₂

Many passively ingested gastrointestinal nematodes have a motile environmental iL3 stage that invades hosts after being swallowed. For example, iL3s of the ruminant parasite Haemonchus contortus inhabit the soil and infect after being swallowed by grazing animals (O'Connor et al., Reference O'Connor, Walkden-Brown and Kahn2006). After entering a host, the nematodes exsheath in the rumen and travel to the abomasum, where they develop into parasitic adults (Laing et al., Reference Laing, Kikuchi, Martinelli, Tsai, Beech, Redman, Holroyd, Bartley, Beasley, Britton, Curran, Devaney, Gilabert, Hunt, Jackson, Johnston, Kryukov, Li, Morrison, Reid, Sargison, Saunders, Wasmuth, Wolstenholme, Berriman, Gilleard and Cotton2013). Similarly, the murine gastrointestinal parasite Heligmosomoides polygyrus has an iL3 stage that can infect mice either from feces during coprophagy or from the fur during grooming (Hernandez and Sukhdeo, Reference Hernandez and Sukhdeo1995). Despite their passive route of infection, both Haemonchus contortus and Heligmosomoides polygyrus actively migrate towards host-associated sensory cues. This suggests that these species use host-associated cues to position themselves in the vicinity of potential hosts, where they are more likely to be ingested (Hernandez and Sukhdeo, Reference Hernandez and Sukhdeo1995; Castelletto et al., Reference Castelletto, Gang, Okubo, Tselikova, Nolan, Platzer, Lok and Hallem2014; Ruiz et al., Reference Ruiz, Castelletto, Gang and Hallem2017; Bryant et al., Reference Bryant, Ruiz, Gang, Castelletto, Lopez and Hallem2018).

Examination of the CO₂-evoked behaviours of Haemonchus contortus and Heligmosomoides polygyrus revealed that both species show experience-dependent responses to CO₂ (Fig. 1C) (Castelletto et al., Reference Castelletto, Gang, Okubo, Tselikova, Nolan, Platzer, Lok and Hallem2014; Ruiz et al., Reference Ruiz, Castelletto, Gang and Hallem2017). In the case of Heligmosomoides polygyrus, iL3s extracted directly from feces are repelled by CO₂, while iL3s that have been removed from feces for multiple days – a condition designed to mimic the soil environment of iL3s – are attracted to CO₂ (Ruiz et al., Reference Ruiz, Castelletto, Gang and Hallem2017). This shift in CO₂ preference appears to occur as a result of the drop in ambient CO₂ levels experienced by the iL3s after they migrate off feces, since cultivating iL3s off feces under high CO₂ conditions (2.5% CO₂) prevents the behavioural switch. The initial repulsion from CO₂ experienced by Heligmosomoides polygyrus iL3s on feces may enable them to disperse off of feces and into the environment to host seek. Following a prolonged period without feces, CO₂ attraction may drive them towards new hosts or fresh host feces to increase their chances of host entry through ingestion (Ruiz et al., Reference Ruiz, Castelletto, Gang and Hallem2017).

In the case of Haemonchus contortus, iL3s directly removed from feces are neutral to CO₂, whereas iL3s that have been removed from feces for a week or more are attracted to CO₂ (Castelletto et al., Reference Castelletto, Gang, Okubo, Tselikova, Nolan, Platzer, Lok and Hallem2014; Ruiz et al., Reference Ruiz, Castelletto, Gang and Hallem2017). This experience-dependent shift in CO₂-evoked behaviour may enable the iL3s to migrate towards the mouths of grazing ruminants, whose breath emits high concentrations of CO₂ (Ruiz et al., Reference Ruiz, Castelletto, Gang and Hallem2017). In contrast to Heligmosomoides polygyrus and Haemonchus contortus, the skin-penetrating nematodes Ancylostoma ceylanicum, Strongyloides stercoralis and Strongyloides ratti do not display this flexibility in their behavioural responses to CO₂. Thus, experience-dependent plasticity towards CO₂ may be unique to passively ingested nematodes (Ruiz et al., Reference Ruiz, Castelletto, Gang and Hallem2017). However, skin-penetrating nematodes do show other forms of sensory plasticity, including experience-dependent thermal plasticity and temperature-dependent olfactory plasticity (Lee et al., Reference Lee, Dillman and Hallem2016; Bryant et al., Reference Bryant, Ruiz, Gang, Castelletto, Lopez and Hallem2018). This suggests that experience-dependent responses to CO₂ may not be beneficial for skin-penetrating nematodes, likely because the skin surface of mammals emits only very low levels of CO₂ (Alkalay et al., Reference Alkalay, Suetsugu, Constantine and Stein1971).

The role of CO₂ in mammalian-parasitic nematode development and physiology

In addition to being a robust behavioural cue for parasitic nematodes, CO₂ is also an important regulator of their development and physiology. For example, CO₂ stimulates exsheathment and activation (exit from the developmentally arrested iL3 stage) in passively ingested ruminant parasites such as Haemonchus contortus (Rogers and Sommerville, Reference Rogers and Sommerville1960; Taylor and Whitlock, Reference Taylor and Whitlock1960; Sommerville, Reference Sommerville1964; Bekelaar et al., Reference Bekelaar, Waghorn, Tavendale, McKenzie and Leathwick2018, Reference Bekelaar, Waghorn, Tavendale, McKenzie and Leathwick2019). However, the requirement for CO₂ during exsheathment varies across species. CO₂ is an absolute requirement for the exsheathment of Haemonchus contortus iL3s, whereas CO₂ enhances but is not required for exsheathment in other passively ingested abomasal nematodes (Bekelaar et al., Reference Bekelaar, Waghorn, Tavendale, McKenzie and Leathwick2018). In the dog hookworm Ancylostoma caninum, CO₂ is not required for activation but results in a slight increase in the rate of activation (Hawdon and Schad, Reference Hawdon and Schad1990).

The role of CO₂ is not limited to exsheathment and activation. CO₂, in combination with O₂, also regulates the development of Strongyloides ratti into either free-living adults or iL3s (Taylor and Weinstein, Reference Taylor and Weinstein1990). In addition, CO₂ stimulates egg hatching in the giant roundworm Ascaris lumbricoides, a human-parasitic species that infects when eggs containing developmentally arrested infective larvae are swallowed by hosts as a result of fecal–oral contamination (Fairbairn, Reference Fairbairn1961; Dold and Holland, Reference Dold and Holland2011). Finally, CO₂ is required for the in vitro development of parasitic larvae in the pig roundworm Ascaris suum (Douvres and Urban, Reference Douvres and Urban1983). Thus, CO₂ influences both behaviour and development in many if not all mammalian-parasitic nematode species.

Introduction to entomopathogenic nematodes

Entomopathogenic nematodes (EPNs) are parasites that infect and kill insects (Dillman and Sternberg, Reference Dillman and Sternberg2012). They are considered beneficial for humans due to their role as biological agents for pest control, and are likely also important for maintaining balanced ecosystems in nature. EPNs of the genera Heterorhabditis and Steinernema have been successfully employed commercially against insect agricultural pests (Liu et al., Reference Liu, Poinar and Berry2000; Grewal et al., Reference Grewal, Ehlers and Shapiro-Ilan2005; Dillman and Sternberg, Reference Dillman and Sternberg2012; Labaude and Griffin, Reference Labaude and Griffin2018). The geographical distribution of EPNs spans all continents except Antarctica (Hominick, Reference Hominick and Gaugler2002). Some EPNs, such as Steinernema carpocapsae and Heterorhabditis bacteriophora, are generalists that can infect many different insects; in contrast, other EPNs have very narrow host ranges (Peters, Reference Peters1996). For example, the specialist Steinernema scapterisci specifically infects mole crickets, and the specialist Steinernema diaprepesi specifically infects the larval stages of the citrus pest Diaprepes abbreviatus (Nguyen and Smart, Reference Nguyen and Smart1991; Nguyen and Hunt, Reference Nguyen and Hunt2007; Ali et al., Reference Ali, Alborn and Stelinski2010). EPNs infect only as third-stage larvae called infective juveniles (IJs); the IJ stage of EPNs is equivalent to the iL3 stage of mammalian-parasitic nematodes (Dillman et al., Reference Dillman, Chaston, Adams, Ciche, Goodrich-Blair, Stock and Sternberg2012a). IJs enter their insect hosts through a body orifice such as the mouth, spiracles or anus; IJs of some species can also penetrate directly through the cuticle (Bedding and Molyneux, Reference Bedding and Molyneux1982; Kaya and Gaugler, Reference Kaya and Gaugler1993). The IJs then enter the insect haemocoel and release a bacterial symbiont from their intestine (Bedding and Molyneux, Reference Bedding and Molyneux1982; Kaya and Gaugler, Reference Kaya and Gaugler1993). Toxins secreted by the nematode and the bacteria kill the insect, typically within 48 h (Kaya and Gaugler, Reference Kaya and Gaugler1993; Lu et al., Reference Lu, Macchietto, Chang, Barros, Baldwin, Mortazavi and Dillman2017; Chang et al., Reference Chang, Serra, Lu, Mortazavi and Dillman2019). The nematodes then feed on the insect cadaver and complete their parasitic life cycle. The nematodes can cycle through multiple generations in the host cadaver until resources are depleted, at which point new IJs form and disperse into the environment to seek out new hosts (Kaya and Gaugler, Reference Kaya and Gaugler1993).

The role of CO₂ in the host-seeking behaviours of EPNs

The host-seeking strategies of EPNs vary across species. Some species are considered ‘cruisers’ that actively migrate towards stationary hosts, other species are considered ‘ambushers’ that remain relatively stationary and nictate to facilitate attachment to mobile hosts, and still other species use an intermediate strategy (Campbell and Gauger, Reference Campbell and Gauger1993; Lewis, Reference Lewis and Gauger2002; Lewis et al., Reference Lewis, Campbell, Griffin, Kaya and Peters2006). However, both ambushers and cruisers are capable of migrating towards host-emitted chemosensory cues, suggesting that all EPNs engage in chemosensory-driven navigation towards hosts (Schmidt and All, Reference Schmidt and All1979; Pye and Burman, Reference Pye and Burman1981; O'Halloran and Burnell, Reference O'Halloran and Burnell2003; Hallem et al., Reference Hallem, Dillman, Hong, Zhang, Yano, DeMarco and Sternberg2011a; Dillman et al., Reference Dillman, Guillermin, Lee, Kim, Sternberg and Hallem2012b; Castelletto et al., Reference Castelletto, Gang, Okubo, Tselikova, Nolan, Platzer, Lok and Hallem2014; Lee et al., Reference Lee, Dillman and Hallem2016). Some EPNs in the genus Steinernema also engage in a unique jumping behaviour where the IJ stands on its tail and then propels itself into the air, presumably to facilitate host attachment as well as transport to new niches (Campbell and Kaya, Reference Campbell and Kaya1999). Jumping can be stimulated by exposure to host-emitted chemosensory cues (Campbell and Kaya, Reference Campbell and Kaya1999, Reference Campbell and Kaya2000; Hallem et al., Reference Hallem, Dillman, Hong, Zhang, Yano, DeMarco and Sternberg2011a; Dillman et al., Reference Dillman, Guillermin, Lee, Kim, Sternberg and Hallem2012b).

Many EPNs, including Heterorhabitis bacteriophora, Steinernema carpocapsae, Steinernema riobrave, Steinernema scapterisci and Steinernema glaseri, are attracted to CO₂ (Fig. 2A) (Gaugler et al., Reference Gaugler, LeBeck, Nakagaki and Boush1980, Reference Gaugler, Campbell and Gupta1991; Lewis et al., Reference Lewis, Gauger and Harrison1993; Robinson, Reference Robinson1995; Hallem et al., Reference Hallem, Dillman, Hong, Zhang, Yano, DeMarco and Sternberg2011a; Dillman et al., Reference Dillman, Guillermin, Lee, Kim, Sternberg and Hallem2012b; Lee et al., Reference Lee, Dillman and Hallem2016). This group includes both specialists and generalists, and both ambushers and cruisers. In addition, CO₂ stimulates jumping in Steinernema carpocapsae, Steinernema riobrave and Steinernema scapterisci IJs at concentrations as low as 0.08% (approximately two times higher than atmospheric levels), suggesting that jumping is highly sensitive to environmental CO₂ (Hallem et al., Reference Hallem, Dillman, Hong, Zhang, Yano, DeMarco and Sternberg2011a; Dillman et al., Reference Dillman, Guillermin, Lee, Kim, Sternberg and Hallem2012b). Both the attractive responses of EPNs towards the odour of live insect hosts and jumping responses to host odour are decreased when CO₂ is chemically removed using a soda lime filter, illustrating the importance of CO₂ for host seeking (Gaugler et al., Reference Gaugler, Campbell and Gupta1991; Dillman et al., Reference Dillman, Guillermin, Lee, Kim, Sternberg and Hallem2012b). However, the extent to which host attraction is reduced in the absence of CO₂ varies across different EPN–host combinations (Dillman et al., Reference Dillman, Guillermin, Lee, Kim, Sternberg and Hallem2012b). Thus, EPNs use CO₂ in combination with host-specific olfactory cues to migrate towards insects. CO₂ also acts synergistically with plant root volatiles to attract some EPNs to plants infested with insects (Turlings et al., Reference Turlings, Hiltpold and Rasmann2012).

Fig. 2. Behavioural responses of entomopathogenic nematodes (EPNs) to CO₂. (A) Behavioural responses of the infective juveniles (IJs) of various EPN species to CO₂ in a chemotaxis assay (Fig. 1A). All EPNs tested showed attraction to 2.5% CO₂. Data are from Dillman et al. (Reference Dillman, Guillermin, Lee, Kim, Sternberg and Hallem2012b) and Hallem et al. (Reference Hallem, Dillman, Hong, Zhang, Yano, DeMarco and Sternberg2011a). (B) Age-dependent changes in the CO₂ preferences of Steinernema scapterisci IJs. IJs were grown at room temperature (approximately 22 °C) and then incubated at 15 °C until the indicated time points (days post-collection). CO₂ responsiveness changes from repulsion to attraction with age. IJs were tested with 1% CO₂. Data are from Lee et al. (Reference Lee, Dillman and Hallem2016). The graph shows medians and interquartile ranges.

Like some mammalian-parasitic nematodes, some EPNs exhibit plasticity in their olfactory responses to CO₂. For example, the IJs of Steinernema scapterisci are repelled by CO₂ immediately after emergence from the insect host, but their response shifts to robust attraction over the course of weeks (Fig. 2B) (Lee et al., Reference Lee, Dillman and Hallem2016). In addition, the rate at which this shift occurs varies depending on the cultivation temperature of the IJs, with the shift occurring more rapidly in IJs cultured at 15 °C relative to IJs cultured at 25 °C (Lee et al., Reference Lee, Dillman and Hallem2016). The change in CO₂ preference correlates with a corresponding change in responses to host odours, including the odour of its natural host, the mole cricket (Lee et al., Reference Lee, Dillman and Hallem2016). The strong repulsion of newly emerged Steinernema scapterisci IJs to CO₂ and host odours may serve as a dispersal mechanism to drive them towards new niches. The mechanisms that drive the change in CO₂ preference following host emergence in Steinernema scapterisci have not yet been elucidated. An intriguing possibility is that the CO₂ repulsion exhibited by IJs immediately following host emergence could result from the elevated levels of CO₂ experienced inside the decaying insect cadaver, similar to the way in which CO₂ repulsion in Heligmosomoides polygyrus iL3s results from the elevated levels of CO₂ experienced on mammalian feces (Ruiz et al., Reference Ruiz, Castelletto, Gang and Hallem2017). However, additional experiments will be necessary to determine whether the change in CO₂ preference over time (or age) in Steinernema scapterisci is in fact regulated by ambient CO₂ levels.

Introduction to plant-parasitic nematodes

Plant-parasitic nematodes (PPNs) are a major cause of agricultural crop damage throughout the world. It has been estimated that PPNs are responsible for approximately 100 billion dollars of crop loss per year worldwide (Jasmer et al., Reference Jasmer, Goverse and Smant2003; Wrather et al., Reference Wrather, Koenning and Anderson2003). Of over 4100 species of PPNs that have been identified (Decraemer and Hunt, Reference Decraemer, Hunt, Perry and Moens2006), the ones that cause the most severe economic loss are the nematodes that infect the roots of major agricultural crops (Bernard et al., Reference Bernard, Egnin, Bonsi, Shah and Mahamood2017). These PPNs prevent water and nutrient uptake by plant roots, which results in greatly reduced crop quality and yield (Bernard et al., Reference Bernard, Egnin, Bonsi, Shah and Mahamood2017).

Responses of PPNs to CO₂

CO₂ is ubiquitously produced by the roots of plants. Several studies have demonstrated an important role for CO₂ in mediating attraction of PPNs to their host plants. For example, the stem nematode Ditylenchus dipsaci, which infects onion and garlic, migrates towards CO₂ (Klingler, Reference Klingler1972; Viglierchio, Reference Viglierchio1990). Many other PPNs, including species from the genera Ditylenchus, Meloidogyne, Heterodera and Pratylenchus, are also attracted to CO₂ (Johnson and Viglierchio, Reference Johnson and Viglierchio1961; Prot, Reference Prot1980; McCallum and Dusenbery, Reference McCallum and Dusenbery1992; Robinson, Reference Robinson1995). In the case of Meloidogyne incognita, attraction to tomato root volatiles appears to be due to the presence of O₂ and CO₂ in the volatile mix (McCallum and Dusenbery, Reference McCallum and Dusenbery1992). However, a more recent study found that for Meloidogyne hapla, the attractant is not CO₂ itself but rather the low pH environment created by dissolved CO₂ (Wang et al., Reference Wang, Bruening and Williamson2009). In the case of the pine wilt nematode Bursaphelenchus xylophilus, the fourth-stage juveniles (J_IVs) are repelled by CO₂. CO₂ repulsion by J_IVs plays an important role in dispersal from its insect vector, the pine sawyer beetle, into the pine tree (Wu et al., Reference Wu, Wickham, Zhao and Sun2019). Bursaphelenchus xylophilus J_IVs enter the beetle tracheal system, where they are transported by the beetle to new pine trees. As the beetle matures and feeds on the pine tree, CO₂ levels in the beetle tracheal system increase. Once CO₂ levels reach a certain concentration, CO₂ repulsion drives the J_IVs out of the beetle spiracles and into the pine tree (Wu et al., Reference Wu, Wickham, Zhao and Sun2019). Thus, the responses of PPNs to CO₂ vary greatly across species. A better understanding of how other PPNs respond to CO₂ may enable the development of new biocontrol strategies.

Introduction to free-living nematodes

Free-living nematodes are found in a wide range of ecological habitats. These include various types of soil, sediment and organic matter, as well as marine and freshwater environments. Free-living nematodes use a wide variety of sensory cues to navigate their environment in search of food and mates, and to escape from predators and pathogens. CO₂ is universally present in terrestrial and aquatic habitats, and may serve as an important cue for survival and propagation of these nematodes. The most well-studied free-living nematode is the model worm Caenorhabditis elegans. C. elegans is commonly found in microbe-rich environments such as those of fallen rotting fruits (Felix and Duveau, Reference Felix and Duveau2012), where CO₂ is produced as one of many microbial byproducts. Consequently, C. elegans displays several behavioural and physiological responses to CO₂.

Behavioural responses of C. elegans to carbon dioxide

The first studies of CO₂ responsiveness in C. elegans demonstrated that these worms undergo rapid changes in locomotion in response to changes in CO₂ concentrations (Dusenbery, Reference Dusenbery1985). These responses are characterized by an overall decrease in movement and an increase in turning frequency (Dusenbery, Reference Dusenbery1985). A more recent study examining the effects of acute CO₂ exposure found that freely moving well-fed adults reverse rapidly when their head is exposed to high levels of CO₂, indicating that CO₂ is an aversive cue for well-fed C. elegans adults (Hallem and Sternberg, Reference Hallem and Sternberg2008). In addition, well-fed C. elegans adults avoid high CO₂ areas when allowed to migrate along a CO₂ gradient in a CO₂-chemotaxis assay (Fig. 3A) (Bretscher et al., Reference Bretscher, Busch and de Bono2008). For well-fed adults, CO₂ may indicate the presence of potential predators or pathogens, and repulsion from CO₂ may function as an escape mechanism.

Fig. 3. Behavioural responses of C. elegans to CO₂. (A) Responses of wild-type C. elegans adults and dauers to CO₂ in a chemotaxis assay (Fig. 1A). Dauers are developmentally arrested third-stage larvae that are similar to parasitic iL3s and IJs (Hotez et al., Reference Hotez, Hawdon and Schad1993; Viney et al., Reference Viney, Thompson and Crook2005; Crook, Reference Crook2014). Animals were either well-fed adults cultivated at ambient CO₂, well-fed adults cultivated at high CO₂, starved adults cultivated at ambient CO₂ or dauer larvae cultivated at ambient CO₂. Adults were tested in a 20 min assay; dauer larvae were tested in a 1 h assay. Responses shown are to 2.5% CO₂ (for adults cultivated at high CO₂) or 10% CO₂ (for all other conditions). For the high CO₂ condition, adults were cultivated at 2.5% CO₂ for one generation prior to the assay. For the starvation condition, adults were starved for 3 h prior to the assay. Data are from Guillermin et al. (Reference Guillermin, Carrillo and Hallem2017), Rengarajan et al. (Reference Rengarajan, Yankura, Guillermin, Fung and Hallem2019) and Hallem et al. (Reference Hallem, Dillman, Hong, Zhang, Yano, DeMarco and Sternberg2011a). ****P < 0.0001, one-way ANOVA with Dunnett's post-test. The graph shows medians and interquartile ranges. (B) Previously experienced hypoxic conditions modulate CO₂ responsiveness in C. elegans adults. Animals cultivated at low (1%) oxygen for 1 h prior to assays showed decreased CO₂ avoidance. This change is mediated by hypoxia inducible factor-1 (HIF-1), since hif-1 mutants are not affected by prior O₂ exposure. ns, not significant relative to N2 (wild-type); ***P < 0.001 relative to N2; +++P < 0.001 relative to N2 exposed to 1% O₂ for 1 h. The N2 (air) condition represents a control condition in which animals were not exposed to a CO₂ gradient. The graph shows means and SEMs. Figure is from Bretscher et al. (Reference Bretscher, Busch and de Bono2008), copyright 2008 National Academy of Sciences. (C) The BAG sensory neurons and GCY-9 are required for CO₂ repulsion in C. elegans. BAG-ablated animals and gcy-9 loss-of-function (lf) mutants are neutral to CO₂. ****P < 0.0001, one-way ANOVA with Dunnett's post-test. The graph shows medians and interquartile ranges. Data are from Carrillo et al. (Reference Carrillo, Guillermin, Rengarajan, Okubo and Hallem2013). (D) Calcium activity in BAG neurons in response to 10% CO₂, as measured using the ratiometric calcium indicator yellow cameleon 3.60 (YC3.60). Green traces show responses of individual neurons; black line shows median response. Data are from Rengarajan et al. (Reference Rengarajan, Yankura, Guillermin, Fung and Hallem2019). (E) A model for experience-dependent modulation of CO₂ responsiveness in C. elegans adults. Under ambient conditions, CO₂ repulsion is mediated by activation of the RIA and RIG interneurons and inhibition of the AIY interneurons, and by both dopamine and neuropeptide signalling involving the neuropeptide gene nlp-1. CO₂ attraction in animals cultivated at high CO₂ is mediated by activation of AIY, inhibition of RIA and silencing of RIG, as well as neuropeptide signalling involving the neuropeptide gene flp-16. Finally, CO₂ attraction in starved adults is mediated by silencing of RIG and by a change in AIY responses such that activation and inhibition are observed with approximately equal frequency. Octopamine signalling and neuropeptide signalling via the neuropeptide genes nlp-1 and flp-16 also promote CO₂ attraction during starvation. Blue = excitatory activity, orange = inhibitory activity, grey = silencing of activity. Figure is adapted from Guillermin et al. (Reference Guillermin, Carrillo and Hallem2017) and Rengarajan et al. (Reference Rengarajan, Yankura, Guillermin, Fung and Hallem2019).

As is the case for some parasitic nematodes, CO₂ responsiveness in C. elegans is subject to experience-dependent plasticity. In the case of C. elegans, one of the factors that influences CO₂ responsiveness is the worm's nutritional status. Depriving adults of food suppresses CO₂ avoidance behaviour (Bretscher et al., Reference Bretscher, Busch and de Bono2008; Hallem and Sternberg, Reference Hallem and Sternberg2008). Moreover, as C. elegans adults are starved, CO₂ response shifts from repulsion to attraction (Fig. 3A) (Rengarajan et al., Reference Rengarajan, Yankura, Guillermin, Fung and Hallem2019). The effects of starvation can be reversed when the animals are re-exposed to food (Rengarajan et al., Reference Rengarajan, Yankura, Guillermin, Fung and Hallem2019). The shift from CO₂ repulsion to CO₂ attraction in starved worms may be beneficial for survival, since starved animals must find food in order to survive and bacterial food emits CO₂. CO₂ attraction by starved animals may also reflect an increased tolerance for risk taking; CO₂ attraction is an inherently risky behaviour, since both predators and pathogens of C. elegans emit CO₂ (Felix and Duveau, Reference Felix and Duveau2012; Brandt and Ringstad, Reference Brandt and Ringstad2015; Schulenburg and Felix, Reference Schulenburg and Felix2017).

CO₂ responsiveness in C. elegans adults is also modulated by recently experienced environmental CO₂ and oxygen (O₂) levels, as well as immediate O₂ context. For example, animals exposed to elevated CO₂ levels (2.5% CO₂) become robustly attracted to CO₂ over the course of hours in a reversible manner (Fig. 3A) (Guillermin et al., Reference Guillermin, Carrillo and Hallem2017). Prior exposure to low O₂ levels also suppresses CO₂ avoidance in adults, an effect that depends on the hypoxia inducible factor gene hif-1 (Fig. 3B) (Bretscher et al., Reference Bretscher, Busch and de Bono2008). In addition, whether animals have been pre-exposed to low O₂ affects their responsiveness to CO₂ stimuli under some conditions (Fenk and de Bono, Reference Fenk and de Bono2017). CO₂ responsiveness is also modulated by ambient O₂ levels such that animals assayed under low O₂ conditions are more strongly repelled by CO₂ than animals assayed under high O₂ conditions (Carrillo et al., Reference Carrillo, Guillermin, Rengarajan, Okubo and Hallem2013; Kodama-Namba et al., Reference Kodama-Namba, Fenk, Bretscher, Gross, Busch and de Bono2013). C. elegans prefers O₂ concentrations below atmospheric (Gray et al., Reference Gray, Karow, Lu, Chang, Chang, Ellis, Marletta and Bargmann2004); however, when exposed to opposing gradients of O₂ and CO₂, the avoidance response to high CO₂ dominates over the avoidance response to high O₂ (Bretscher et al., Reference Bretscher, Busch and de Bono2008). In nature, C. elegans is found in rotting vegetation (Felix and Braendle, Reference Felix and Braendle2010; Schulenburg and Felix, Reference Schulenburg and Felix2017), where both O₂ and CO₂ concentrations fluctuate. Moreover, both O₂- and CO₂-sensing pathways control foraging behaviour (Bendesky et al., Reference Bendesky, Tsunozaki, Rockman, Kruglyak and Bargmann2011; Milward et al., Reference Milward, Busch, Murphy, de Bono and Olofsson2011; Juozaityte et al., Reference Juozaityte, Pladevall-Morera, Podolska, Norgaard, Neumann and Pocock2017). Thus, the interplay between O₂- and CO₂-evoked behaviours likely contributes to the ability of C. elegans to navigate the complex organic environments it inhabits.

CO₂ responsiveness is also modulated by the presence or absence of food, and prior temperature experience (Bretscher et al., Reference Bretscher, Busch and de Bono2008, Reference Bretscher, Kodama-Namba, Busch, Murphy, Soltesz, Laurent and de Bono2011; Kodama-Namba et al., Reference Kodama-Namba, Fenk, Bretscher, Gross, Busch and de Bono2013). In the case of temperature-dependent modulation of CO₂ responsiveness, animals cultivated at 22 °C show enhanced repulsion to 1% CO₂ when assayed at 15 °C compared with 22 °C (Kodama-Namba et al., Reference Kodama-Namba, Fenk, Bretscher, Gross, Busch and de Bono2013), suggesting an interaction between recent temperature experience and CO₂ sensing. The ecological significance of this interaction is not yet clear, but it suggests that CO₂ preferences may vary on a diurnal cycle as the ambient temperature fluctuates. Thus, CO₂-evoked behaviours are regulated by multiple sensory modalities, resulting in both context-dependent and experience-dependent responses to CO₂.

CO₂ can also elicit behavioural changes in C. elegans that are independent of changes in locomotion. For example, exposing C. elegans to high CO₂ levels (5% CO₂) inhibits egg-laying behaviour, at least transiently (Fenk and de Bono, Reference Fenk and de Bono2015). Modulation of egg-laying behaviour by high CO₂ levels may prevent animals from exposing their progeny to unfamiliar environmental conditions. Feeding behaviour is also altered by CO₂ such that brief exposure to high CO₂ levels causes an acute reduction in pharyngeal pumping (Sharabi et al., Reference Sharabi, Hurwitz, Simon, Beitel, Morimoto, Rechavi, Sznajder and Gruenbaum2009). Well-fed animals stop pharyngeal pumping when exposed to a lower concentration of CO₂ than starved animals, suggesting that this behavioural change is dependent on the nutritional state of the animal (Sharabi et al., Reference Sharabi, Hurwitz, Simon, Beitel, Morimoto, Rechavi, Sznajder and Gruenbaum2009).

Finally, C. elegans exhibits life-stage-specific responses to CO₂. C. elegans dauers, which are developmentally arrested third-stage larvae that are similar to the iL3 and IJ stages of parasitic nematodes (Hotez et al., Reference Hotez, Hawdon and Schad1993; Viney et al., Reference Viney, Thompson and Crook2005; Crook, Reference Crook2014), are attracted to CO₂ (Fig. 3A) (Hallem et al., Reference Hallem, Dillman, Hong, Zhang, Yano, DeMarco and Sternberg2011a). Dauer larvae form when food is limited or environmental conditions are otherwise unfavourable (Hu, Reference Hu2007), and CO₂ attraction by dauers may enable them to locate bacterial food sources. In addition, dauers associate with invertebrates such as slugs, snails and isopods, which act as carriers that facilitate their dispersal to new niches (Baird, Reference Baird1999; Caswell-Chen et al., Reference Caswell-Chen, Chen, Lewis, Douhan, Nadler and Carey2005; Lee et al., Reference Lee, Choi, Lee, Kim, Hwang, Kim, Park, Paik and Lee2012; Schulenburg and Felix, Reference Schulenburg and Felix2017). Thus, CO₂ attraction may also enable dauers to locate and associate with invertebrate carriers.

Effects of CO₂ on C. elegans development and physiology

High levels of CO₂ can also elicit physiological changes in C. elegans. Cultivating animals at CO₂ levels above 9% CO₂ decreases brood size and significantly slows the rate of embryonic development (Sharabi et al., Reference Sharabi, Hurwitz, Simon, Beitel, Morimoto, Rechavi, Sznajder and Gruenbaum2009). This developmental delay is not caused by any obvious reduction in the overall health of the embryos or the adults that develop from these embryos. In addition, chronic exposure (>4 days) to 19% CO₂ causes defects in overall body muscle morphology, resulting in long-lasting movement defects (Sharabi et al., Reference Sharabi, Hurwitz, Simon, Beitel, Morimoto, Rechavi, Sznajder and Gruenbaum2009). Nevertheless, cultivation at 19% CO₂ extends lifespan, an effect that appears to be independent of the inhibitory effect of CO₂ on egg laying (Sharabi et al., Reference Sharabi, Hurwitz, Simon, Beitel, Morimoto, Rechavi, Sznajder and Gruenbaum2009). The lifespan extension induced by exposure to high CO₂ also may be independent of the dietary restriction pathway, since eat-2 mutants, which show reduced pharyngeal pumping, show increased lifespan at high CO₂ (Sharabi et al., Reference Sharabi, Hurwitz, Simon, Beitel, Morimoto, Rechavi, Sznajder and Gruenbaum2009). Thus, CO₂ has diverse effects on C. elegans development, physiology and behaviour.

Behavioural and physiological responses of other free-living nematodes to CO₂

Other free-living nematodes show diverse responses to CO₂. For example, the free-living marine nematode Adoncholaimus thalassophygas is attracted to CO₂ (Riemann and Schrage, Reference Riemann and Schrage1988). This effect is not due to a general decrease in the pH of the medium, since the addition of hydrochloric acid did not elicit a similar attractive response (Riemann and Schrage, Reference Riemann and Schrage1988). CO₂ is emitted from bacteria present in decaying carcasses found in sediments and may act as a food signal for these nematodes. The necromenic nematode Pristionchus pacificus, which represents an evolutionary intermediate between free-living and parasitic nematodes, displays acute CO₂ avoidance (Hallem and Sternberg, Reference Hallem and Sternberg2008). In contrast, the free-living nematodes Caenorhabditis briggsae, Caenorhabditis angaria and Panagrellus redivivus do not respond to CO₂ in acute avoidance assays (Hallem and Sternberg, Reference Hallem and Sternberg2008). However, ambient CO₂ concentrations play a prominent role in regulating reproduction in Panagrellus redivivus. Under low O₂ conditions, the brood size of Panagrellus redivivus increases in response to an increase in CO₂ concentration from 0 to 5% (Hansen and Buecher, Reference Hansen and Buecher1970). The effects of CO₂ on nematode physiology also vary greatly across species. At the extreme, nematodes have been isolated from volcanic gas vents, where CO₂ levels can reach 100%; these nematodes can survive under 100% CO₂ conditions for at least 5 days (Pilz and Hohberg, Reference Pilz and Hohberg2015). Thus, CO₂ has species-specific effects on nematode behaviour and physiology.

Cellular mechanisms of CO₂ responsiveness in C. elegans

The primary CO₂-sensing neurons in C. elegans are the paired BAG neurons in the head. Ablation of the BAG neurons abolishes both CO₂ avoidance in adults (Hallem and Sternberg, Reference Hallem and Sternberg2008) and CO₂ attraction in dauers (Fig. 3C) (Hallem et al., Reference Hallem, Dillman, Hong, Zhang, Yano, DeMarco and Sternberg2011a). Exposing animals to varying concentrations of CO₂ produces dose-dependent calcium activity in the BAG neurons in vivo (Fig. 3D) (Hallem et al., Reference Hallem, Spencer, McWhirter, Zeller, Henz, Ratsch, Miller, Horvitz, Sternberg and Ringstad2011b). In addition, isolated BAG neurons derived from C. elegans embryos respond to CO₂ in vitro, suggesting that BAG neurons are intrinsically sensitive to CO₂ (Smith et al., Reference Smith, Martinez-Velazquez and Ringstad2013). The calcium responses in isolated BAG neurons in culture are independent of carbonic anhydrase activity, indicating that these neurons can sense molecular CO₂. Moreover, a majority of isolated BAG neurons do not respond to pH, although responses to low pH are detectable in some isolated BAG neurons (Smith et al., Reference Smith, Martinez-Velazquez and Ringstad2013). However, the role of BAG neurons is not limited to promoting CO₂ responses. The BAG neurons also sense O₂ (Zimmer et al., Reference Zimmer, Gray, Pokala, Chang, Karow, Marletta, Hudson, Morton, Chronis and Bargmann2009). Furthermore, they play a role in establishing food odour preferences and in foraging behaviour (Harris et al., Reference Harris, Shen, Ha, Donato, Wallis, Zhang and Zhang2014; Juozaityte et al., Reference Juozaityte, Pladevall-Morera, Podolska, Norgaard, Neumann and Pocock2017). In addition to BAG neurons, other sensory neurons including ASE, AFD, AWC, ASJ, ASK, ASH and ADL also exhibit CO₂-evoked calcium activity and contribute to CO₂ responsiveness (Bretscher et al., Reference Bretscher, Kodama-Namba, Busch, Murphy, Soltesz, Laurent and de Bono2011; Fenk and de Bono, Reference Fenk and de Bono2015). The interneurons AIY, RIG, RIA and AIZ act downstream of BAG neurons to mediate CO₂-evoked behaviour (Kodama-Namba et al., Reference Kodama-Namba, Fenk, Bretscher, Gross, Busch and de Bono2013; Guillermin et al., Reference Guillermin, Carrillo and Hallem2017). The AIA interneurons also show CO₂-evoked activity and are involved in CO₂ responsiveness (Fenk and de Bono, Reference Fenk and de Bono2015). In well-fed adults, CO₂ repulsion is correlated with activation of RIA, RIG and AIZ, and inhibition of AIY (Fig. 3E) (Guillermin et al., Reference Guillermin, Carrillo and Hallem2017).

Molecular mechanisms of CO₂ responsiveness in C. elegans

The detection of CO₂ by the BAG neurons requires the receptor-type guanylate cyclase GCY-9. Animals with loss-of-function mutations in gcy-9 are insensitive to CO₂ in behavioural assays (Fig. 3C) (Hallem et al., Reference Hallem, Spencer, McWhirter, Zeller, Henz, Ratsch, Miller, Horvitz, Sternberg and Ringstad2011b). Moreover, CO₂-evoked calcium transients in the BAG neurons require GCY-9, and ectopic expression of GCY-9 confers CO₂ sensitivity to other sensory neurons (Hallem et al., Reference Hallem, Spencer, McWhirter, Zeller, Henz, Ratsch, Miller, Horvitz, Sternberg and Ringstad2011b; Brandt et al., Reference Brandt, Aziz-Zaman, Juozaityte, Martinez-Velazquez, Petersen, Pocock and Ringstad2012; Carrillo et al., Reference Carrillo, Guillermin, Rengarajan, Okubo and Hallem2013). The expression of GCY-9 in the BAG neurons requires the E26 transformation-specific (ETS)-domain transcription factor ETS-5, and ets-5 mutants fail to avoid CO₂ (Guillermin et al., Reference Guillermin, Castelletto and Hallem2011; Brandt et al., Reference Brandt, Aziz-Zaman, Juozaityte, Martinez-Velazquez, Petersen, Pocock and Ringstad2012). Both ETS-5 and the SoxD transcription factor EGL-13 are also required more generally for normal differentiation of the BAG neurons (Guillermin et al., Reference Guillermin, Castelletto and Hallem2011; Brandt et al., Reference Brandt, Aziz-Zaman, Juozaityte, Martinez-Velazquez, Petersen, Pocock and Ringstad2012; Petersen et al., Reference Petersen, Romanos, Juozaityte, Riveiro, Hum, Traunmuller, Zimmer and Pocock2013). The Toll-like receptor TOL-1 is also required for normal BAG neuron differentiation, and tol-1 mutants are defective in pathogen avoidance behaviour as a result (Brandt and Ringstad, Reference Brandt and Ringstad2015).

In addition to GCY-9, the cGMP signalling pathway that mediates CO₂ detection consists of the cGMP-gated cation channel TAX-2/TAX-4 (Bretscher et al., Reference Bretscher, Busch and de Bono2008; Hallem and Sternberg, Reference Hallem and Sternberg2008). Glutamate signalling and neuropeptide signalling are also required for BAG-mediated responses to CO₂ (Guillermin et al., Reference Guillermin, Carrillo and Hallem2017). BAG neurons are glutamatergic (Serrano-Saiz et al., Reference Serrano-Saiz, Poole, Felton, Zhang, De La Cruz and Hobert2013), and well-fed adults lacking the vesicular glutamate transporter EAT-4 show neutral responses to CO₂ (Guillermin et al., Reference Guillermin, Carrillo and Hallem2017). BAG neurons also release neuropeptides, and well-fed adults lacking the BAG-expressed FMRFamide-like neuropeptide FLP-17 do not respond to CO₂ (Guillermin et al., Reference Guillermin, Carrillo and Hallem2017). In addition, several other signalling molecules, including the calcineurin subunits TAX-6 and CNB-1, the regulator of G-protein signalling RGS-3 and the nuclear hormone receptor NHR-49 are known to regulate CO₂ response (Hallem and Sternberg, Reference Hallem and Sternberg2008). The microRNA mir-791 is also required for the normal CO₂-evoked calcium activity of the BAG neurons (Drexel et al., Reference Drexel, Mahofsky, Latham, Zimmer and Cochella2016).

Mechanisms underlying the context-dependent modulation of CO₂ responsiveness by O₂

The extent to which CO₂ responsiveness is regulated by ambient O₂ levels depends on the neuropeptide Y receptor NPR-1 (McGrath et al., Reference McGrath, Rockman, Zimmer, Jang, Macosko, Kruglyak and Bargmann2009; Carrillo et al., Reference Carrillo, Guillermin, Rengarajan, Okubo and Hallem2013; Kodama-Namba et al., Reference Kodama-Namba, Fenk, Bretscher, Gross, Busch and de Bono2013). The laboratory wild-type (N2) strain of C. elegans contains a gain-of-function mutation in the npr-1 gene that confers CO₂ avoidance on well-fed adults regardless of ambient O₂ levels. However, animals containing loss-of-function (lf) mutations in npr-1 and animals carrying the natural variant of npr-1 avoid CO₂ under low O₂ conditions but do not respond to CO₂ at normal atmospheric O₂ levels (21% O₂) (Carrillo et al., Reference Carrillo, Guillermin, Rengarajan, Okubo and Hallem2013; Kodama-Namba et al., Reference Kodama-Namba, Fenk, Bretscher, Gross, Busch and de Bono2013). The gain-of-function NPR-1 variant in N2 animals suppresses the activity of the O₂-sensing URX neurons to promote CO₂ avoidance regardless of ambient O₂ levels. In animals containing an npr-1(lf) mutation or a natural variant of npr-1, the URX neurons are tonically active under high O₂ conditions and inhibit CO₂ avoidance at high O₂. The RIA interneurons appear to act downstream of URX to partially mediate its effects on the CO₂ circuit (Kodama-Namba et al., Reference Kodama-Namba, Fenk, Bretscher, Gross, Busch and de Bono2013). In addition, the neuroglobin gene glb-5 also acts via the URX neurons to modulate CO₂ responsiveness as a function of ambient O₂ levels (McGrath et al., Reference McGrath, Rockman, Zimmer, Jang, Macosko, Kruglyak and Bargmann2009; Kodama-Namba et al., Reference Kodama-Namba, Fenk, Bretscher, Gross, Busch and de Bono2013).

Mechanisms underlying the experience-dependent modulation of CO₂ responsiveness

The mechanisms underlying experience-dependent modulation of CO₂ responsiveness in C. elegans have been elucidated in some detail. The shift in CO₂ response from repulsion to attraction when animals are moved from low CO₂ to high CO₂ cultivation conditions results from the differential activity of a single set of interneurons downstream of the BAG sensory neurons (Guillermin et al., Reference Guillermin, Carrillo and Hallem2017). In animals that were previously cultivated at low CO₂, CO₂ exposure inhibits the AIY interneurons and activates the RIA and RIG interneurons. In contrast, in animals that have been cultivated at high CO₂, CO₂ exposure activates AIY and inhibits RIA. Moreover, RIG is silenced such that it no longer responds to CO₂ (Fig. 3E). Thus, CO₂ response is not determined by whether an ‘attractive’ or ‘repulsive’ pathway is activated; rather, it is determined by experience-dependent modulation of interneuron activity in a single pathway (Guillermin et al., Reference Guillermin, Carrillo and Hallem2017). A number of neuropeptides also differentially modulate CO₂ responsiveness in animals cultured under high vs low CO₂ conditions (Fig. 3E) (Guillermin et al., Reference Guillermin, Carrillo and Hallem2017).

The shift from CO₂ repulsion to CO₂ attraction that occurs during starvation also arises due to the differential activities of the AIY and RIG interneurons (Rengarajan et al., Reference Rengarajan, Yankura, Guillermin, Fung and Hallem2019). In starved animals, RIG is silenced and AIY shows stochastic responses such that CO₂ evokes activating and inhibiting responses with approximately equal frequency (Fig. 3E). At the molecular level, whether CO₂ is attractive or repulsive is regulated by biogenic amine signalling. Dopamine promotes CO₂ avoidance in well-fed animals by promoting activation of RIG and inhibition of AIY, while octopamine promotes CO₂ attraction in starved animals by promoting activation of AIY (Fig. 3E) (Rengarajan et al., Reference Rengarajan, Yankura, Guillermin, Fung and Hallem2019). Thus, the CO₂ circuit is modulated during starvation by opposing biogenic amine signals. Neuropeptide signalling also regulates CO₂ responsiveness during starvation (Fig. 3E) (Rengarajan et al., Reference Rengarajan, Yankura, Guillermin, Fung and Hallem2019). Finally, CO₂ attraction in dauer larvae is less well understood but is regulated at least in part by neuropeptide signalling (Lee et al., Reference Lee, Shih, Schaedel, Quintero-Cadena, Rogers and Sternberg2017).

Molecular and cellular mechanisms underlying other CO₂-evoked behaviours

Some of the molecular and cellular mechanisms that mediate the effects of CO₂ on other behaviours in C. elegans have also been elucidated. For instance, CO₂-evoked activity in the AWC sensory neurons triggers a cGMP signalling pathway that ultimately inhibits the activity of the HSN neurons, resulting in the inhibition of egg-laying behaviour (Fenk and de Bono, Reference Fenk and de Bono2015). Antagonistic effects of the BAG neurons and the URX neurons regulate lifespan in C. elegans, resulting in increased longevity in BAG-ablated animals (Liu and Cai, Reference Liu and Cai2013). Mutations in the c-Jun N-terminal kinase (JNK) signalling pathway genes jnk-1 and kgb-2 suppress CO₂-induced fertility defects, indicating that JNK signalling may be involved in regulating fertility in response to CO₂ (Vadasz et al., Reference Vadasz, Dada, Briva, Helenius, Sharabi, Welch, Kelly, Grzesik, Budinger, Liu, Seeger, Beitel, Gruenbaum and Sznajder2012).

Unanswered questions regarding CO₂ responsiveness in C. elegans

Although the mechanisms underlying CO₂ responsiveness in C. elegans have been elucidated in appreciable detail, several questions remain unexplored. For example, more information is needed to fully understand how the differential flow of information from BAG neurons to downstream interneurons generates experience-dependent plasticity of CO₂-evoked behaviour. One intriguing possibility is that the BAG neurons express or release different neurotransmitters or neuropeptides in response to CO₂ under varying conditions. Consistent with this possibility, the BAG neurons modulate the expression of FLP-19 neuropeptides as a function of their CO₂-evoked activity (Rojo Romanos et al., Reference Rojo Romanos, Petersen and Pocock2017). In addition, the interneurons that act downstream of other CO₂-sensing neurons have not been identified. Finally, the CO₂ microcircuit that drives CO₂ attraction in dauers remains poorly understood, although it appears to involve dauer-specific, gap-junction-mediated signalling between the BAG neurons and the downstream AIB interneurons (Bhattacharya et al., Reference Bhattacharya, Aghayeva, Berghoff and Hobert2019). In future studies, it will be interesting to determine whether the same set of neurons or a distinct set of neurons promotes CO₂ attraction in dauers. A better understanding of the neural circuits and signalling pathways that regulate CO₂ responsiveness as a function of experience, context and life stage will provide important insights into how a single sensory cue can give rise to diverse behavioural responses in an ethologically-appropriate manner.

Mechanisms underlying CO₂ responsiveness in other nematodes

The anatomy and function of nematode sensory neurons are generally conserved across species (Ashton et al., Reference Ashton, Bhopale, Fine and Schad1995, Reference Ashton, Li and Schad1999; Lopez et al., Reference Lopez, Boston, Ashton and Schad2000; Li et al., Reference Li, Ashton, Gamble and Schad2000a, Reference Li, Zhu, Boston, Ashton, Gamble and Schad2000b, Reference Li, Zhu, Ashton, Gamble and Schad2001; Bhopale et al., Reference Bhopale, Kupprion, Ashton, Boston and Schad2001; Forbes et al., Reference Forbes, Ashton, Boston, Zhu and Schad2004; Ketschek et al., Reference Ketschek, Joseph, Boston, Ashton and Schad2004; Ashton et al., Reference Ashton, Zhu, Boston, Lok and Schad2007; Bumbarger et al., Reference Bumbarger, Crum, Ellisman and Baldwin2007; Srinivasan et al., Reference Srinivasan, Durak and Sternberg2008; Bumbarger et al., Reference Bumbarger, Wijeratne, Carter, Crum, Ellisman and Baldwin2009; Zhu et al., Reference Zhu, Li, Nolan, Schad and Lok2011; Hallem et al., Reference Hallem, Dillman, Hong, Zhang, Yano, DeMarco and Sternberg2011a), making it possible to use knowledge of CO₂ responsiveness in C. elegans as a starting point for launching investigations into the mechanisms of CO₂ responsiveness in parasitic nematodes. In the case of both the necromenic nematode Pristionchus pacificus and the EPNs Heterorhabditis bacteriophora and Steinernema carpocapsae, BAG neurons were identified on the basis of conserved neuroanatomical position and shown to be required for behavioural responses to CO₂ by laser ablation analyses (Hallem and Sternberg, Reference Hallem and Sternberg2008; Hallem et al., Reference Hallem, Dillman, Hong, Zhang, Yano, DeMarco and Sternberg2011a). BAG-ablated Pristionchus pacificus adults do not show acute CO₂ avoidance, and BAG-ablated Heterorhabditis bacteriophora and Steinernema carpocapsae IJs do not show CO₂ attraction (Hallem et al., Reference Hallem, Dillman, Hong, Zhang, Yano, DeMarco and Sternberg2011a). In addition, CO₂-evoked jumping behaviour in Steinernema carpocapsae requires the BAG neurons (Hallem et al., Reference Hallem, Dillman, Hong, Zhang, Yano, DeMarco and Sternberg2011a). Thus, the neural circuits that mediate CO₂ response are at least partly conserved across nematode species. However, the interneurons that operate downstream of BAG neurons to mediate CO₂ responsiveness in other nematode species have not yet been identified. Moreover, nothing is currently known about the neural circuits and molecular signals that promote CO₂ responsiveness in mammalian-parasitic nematodes. In future studies, it will also be interesting to determine whether similar or distinct mechanisms operate in C. elegans and parasitic nematodes to modulate CO₂ responses depending on context, previous experience or life stage.