Changes in the frequency of facultative endosymbionts in insect populations: overview and applications

Abstract

Many insect endosymbionts are facultative from the host perspective, and their population frequencies across time and space will depend on their transmission fidelity and effects on host fitness. These effects and transmission rates in turn depend on the environmental and host genetic contexts where the endosymbionts occur. Endosymbionts like Wolbachia and Cardinium affect host reproduction to produce transient or persistent presence/absence polymorphisms, while other endosymbionts like Regiella and Hamiltonella persist through providing host fitness benefits and transmitting horizontally. Evolutionary changes in hosts and endosymbionts affect these impacts and endosymbiont polymorphisms in host populations and host sexes. We review this diversity of endosymbiont-host interactions and their influence on the usefulness of endosymbionts for applied strategies. Current strategies focus on endosymbionts driving useful traits to fixation (particularly Wolbachia suppression of arbovirus transmission by mosquitoes) or endosymbionts suppressing populations due to infected males sterilising females. Transinfected endosymbionts sourced from one species and microinjected into another have proven effective in these Wolbachia-mosquito strategies. Novel strategies involving transinfected Rickettsiella, Regiella and Wolbachia may decrease the impacts of pest invertebrates by suppressing pest numbers, reducing the capacity of vector hosts to transmit plant viral diseases or bolstering the effectiveness of natural enemies. Because many endosymbionts are already present in the environment, their applied use raises fewer safety concerns when compared to genetic modification, as supported by more than 13 years of field experiences with Wolbachia in mosquitoes that have not raised major concerns.

1. Introduction

Endosymbiotic bacteria are common in insects, mites and other invertebrates, and are often regarded as essential for host function, particularly through providing nutrition including vitamins and amino acids (Baumann 2005; Douglas 1998). Primary endosymbionts are typically present in almost all conspecifics, with their obligate contributions to host fitness evident in their reduced genomes and metabolic integration with hosts (McCutcheon & Moran 2007; Sloan et al. 2014). Secondary endosymbionts are not necessarily required by hosts (i.e. facultative) and often persist in a polymorphic state (i.e. host individuals with and without the endosymbiont). Examples of obligate relationships in insects include Portiera in whiteflies (Baumann et al. 2004), Carsonella in psyllids (Morrow et al. 2017), Candidatus Moranella and Tremblaya in mealybugs (Zientz et al. 2004) and Buchnera in aphids (Douglas 1998) though the latter may be replaced by Geopemphigus (Chong & Moran 2018). Removal of these endosymbionts typically results in dramatic effects on fitness: for instance, removal of Buchnera with antibiotics decreases host survival and reproduction (e.g. Koga et al. 2007). These obligate associations can continue to evolve over time and involve multiple bacterial groups. For instance, in Lachninae aphids, associations between Serratia and Buchnera have evolved multiple times with Serratia initially providing riboflavin and peptidoglycan pathways lost from Buchnera, followed by later provisioning of additional amino acid resources (Monnin et al. 2020). Bacterial endosymbiont evolution includes patterns of degradation and loss of function with eventual replacement possible. For example, yeast-like fungal symbionts have replaced Buchnera in about 20 aphids and fungal symbionts closely related to parasitic Ophiocordyceps have replaced Hodgkinia bacterial endosymbionts in several cicadas (Fukatsu & Ishikawa 1996; Hongoh & Ishikawa 2000; Matsuura et al. 2018).

The frequency of facultative secondary endosymbionts often varies across related species, populations and even seasonally within populations. The most well-studied of these endosymbionts are strains of Wolbachia alphaproteobacteria, although these can also function as primary endosymbionts in cases where Wolbachia are present in all conspecifics. In this case, their removal can drastically reduce fitness as observed in bed bugs (Nikoh et al. 2014), providing treatment options. However, removal of Wolbachia strains that are polymorphic in populations often has little impact on fitness, as in the case of some Drosophila infections (e.g. Hoffmann et al. 1996). Host fitness can even be improved in some species following Wolbachia removal (e.g. Li et al. 2019).

Many other endosymbionts like Spiroplasma, Regiella, and Hamiltonella are typically polymorphic in invertebrate populations. These endosymbionts often change in frequency quite rapidly in response to environmental factors impacting transmission, host traits and interactions with other microbes which form the main points of discussion in this paper. We provide an overview of these factors that may also impact the density of the endosymbionts in hosts. This is followed by a brief discussion of how these facultative polymorphic endosymbionts have been used in applied settings and particularly how these uses could be expanded. Some potential applications build on efforts to use Wolbachia pathogen blocking in the control of human diseases and the impact of Wolbachia on mating incompatibility which could be expanded from mosquitoes to agricultural pests, as discussed elsewhere and the impact of Wolbachia on mating incompatibility which could be expanded from mosquitoes to agricultural pests, as discussed elsewhere (Gong et al. 2023; Zindel et al. 2011). We discuss wider applications covering a diverse range of situations and organisms including clonal pests. We conclude that characterizing endosymbiont transmission and effects on hosts in the context of environmental and genetic interactions is critical for optimizing novel applications.

2. Nature of endosymbiont polymorphisms

Primary endosymbionts fixed in populations are typically viewed as mutualists that benefit their hosts (Table 1(i)). A mutualistic association is also expected when specialist host cells such as the bacteriocytes of aphids and whiteflies, and the antennal glands of beewolves (Goettler et al. 2007), provide tissues with a high density of endosymbionts in a host. Bacteriocytes show specialized host transcription patterns likely to aid the endosymbiont (Braendle et al. 2003) which may reflect a long history of coevolution. This period follows initial endosymbiont establishment (McCutcheon et al. 2019), which may potentially occur during host speciation (Gerth & Bleidorn 2016; Raychoudhury et al. 2009), horizontally via plant and animal vectors (Ahmed et al. 2015; Li et al. 2017; Zhao et al. 2021), or through host hybridization and introgression (Cooper et al. 2019; Turelli et al. 2018). Regardless, all currently fixed (or nearly fixed) endosymbionts begin as facultative, polymorphic infections.

Table 1.

Overview of factors influencing or potentially influencing endosymbiont frequencies in host populations.

Characteristics of endosymbionts/environments	Nature of interaction	Dynamics	Examples	Factors influencing endosymbiont dynamics	Example references
(i) Nutritional mutualists	Endosymbionts provide resources often essential to host	Fixed or near fixation of endosymbiont	Buchnera and Portiera; removal is typically fatal to host	Essential nutrition provided to host means poor survival in absence of endosymbiont	Koga et al. 2007; Douglas 1998; Baumann et al 2004
(ii) Modification of host reproduction that assists spread	Strong reproductive incompatibility caused by endosymbiont, also parthenogenesis and male killing	For parthenogenesis and incompatibility, endosymbiont can drive endosymbiont to fixation, but population is often polymorphic (also for male killer)	Spreading natural Wolbachia in Drosophila, spreading transinfected Wolbachia in Aedes mosquitoes, stable polymorphic Spiroplasma male killers	Strength of reproductive effects compared to other effects on host and transmission, evolutionary history (e.g. loss of functional males in case of parthenogenesis)	Huigens & Stouthamer 2003; Hoffmann & Turelli 1997; Gong et al. 2020; 2023; Hoffmann et al 2011; Ballinger & Perlman 2019; Zhang et al. 2016
(iii) Environmental stresses on endosymbionts and their transmission	Some conditions lead to loss of endosymbiont or reduce its transmission fidelity	Depends on severity and incidence of stressors	Temperature, quality of nutrition, antibiotics and other chemicals in environment	Nature of environmental stress, evolutionary history (e.g. adaptation to stressful conditions)	Hoffmann et al. 2015; Van Opijen & Breeuwer 1999; Liu et al. 2019; Ross et al. 2019a; Hague et al. 2022; Martins et al. 2023; Serbus et al. 2015; Koga et al. 2007
(iv) Environment-mediated horizontal transmission	Transfer of endosymbiont to uninfected hosts through plant tissue or (more rarely) parasitoids, direct contact during development	Depends on whether endosymbiont can survive in plant tissue or outside of host body	Rickettsiella transmission in aphids, Wolbachia transmission in whitefly	Transient transmission if endosymbiont density is low in new host, depending on ability of endosymbiont to establish, multiply, & influence host fitness	Gehrer & Vorburger 2012; Heath et al. 1999; Gu et al. 2023; Giorgini et al. 2010; Li et al. 2017
(v) Environment-mediated effects on host fitness (abiotic)	Changes in endosymbiont frequencies related to environmental conditions across space and time	Geographically or seasonally affected polymorphisms driven by changing host fitness effects	Decrease in Wolbachia frequencies at high latitudes in Drosophila, Lepidoptera, seasonal changes in frequency of Regiella in aphids	Nature of abiotic conditions influencing fitness of infected hosts, interactions with host adaptation	Kriesner et al. 2016; Zhao et al. 2021; Liu et al. 2019
(vi) Environment-mediated effects on host fitness (biotic)	Changes in endosymbiont frequencies related to variation in biotic factors including host plant, natural enemies, nutrition	Driven by occurrence of biotic factors that influence endosymbiont fitness effects such as protection from natural enemies	Production of pederin by PPE endosymbionts in rove beetles to protect against predators but at a fitness cost; Spiroplasma protection of insects, against parasitoids, nematodes, and fungi; Hamiltonella protection against parasitoids in aphids	As for (v)	Fang et al. 2023; Ballinger & Perlman 2019; Scarborough et al. 2005; Vorburger 2018
(vii) Evolutionary changes in host and variation in host background	Host genetic polymorphisms affecting endosymbiont persistence (e.g. suppressor evolution), expression of endosymbiont-associated phenotypes	Depends on evolutionary history of host and host fitness consequences as well as impact on endosymbiont transmission, likely frequency dependent	Expression of male killing by Wolbachia in Lepidoptera and Drosophila depends on suppression genes in host, population background differences for fitness effects of Rickettsia in whiteflies	Depends on nature of modified endosymbiont effect, for instance suppression of male killing could lead to host population persistence where mating is required for reproduction, and limited impact on endosymbiont frequency	Hornett et al. 2009; Richardson et al. 2023; Cass et al. 2016
(viii) Clonal associations with endosymbionts	Vertically transmitted endosymbionts become associated with specific clones in parthenogenetic/cyclical parthenogenetic species like aphids	Depends on the clone initially colonized by the endosymbiont and subsequent rates of horizontal transfer, clone-fitness interactions	Incidence of Regiella depends on fitness of clone of Sitobion avenae aphid where it occurs	Selection on clones and endosymbiont effects will influence endosymbiont dynamics	Zepeda-Paulo et al. 2017
(ix) Sex-based polymorphisms of endosymbiont presence/density	Endosymbiont density is reduced in males relative to females and may become undetectable in males as they age	Evolutionary loss of endosymbionts in males but not females due to lack of direct male-based selection	Low incidence of wAlbA Wolbachia in male Aedes albopictus mosquitoes, and Wolbachia absence in male Drosophila pseudotakahashii	Depends on whether male phenotypic effect of endosymbiont (like CI) is still expressed when endosymbiont is absent in males	Tortosa et al. 2010; Yang et al. 2022; Richardson et al. 2019
(x) Other evolutionary changes in endosymbionts	Evolution of increased maternal transmission fidelity, mutualism, and other traits speeding spread of endosymbiont	Depends on evolutionary history of endosymbiont-host interactions: for instance, endosymbiont may need to be beneficial to initially spread in new host	Evolved changes in endosymbiont effects as seen in Wolbachia impact on fitness of D. simulans, evolution of metabolic integration as observed for obligate association	Depends on whether endosymbiont can evolve to further benefit host and improve vertical and horizontal transmission	Weeks et al. 2007; McCutcheon & Moran 2007; Sloan et al. 2014
(xi) Presence of other microbes in same host	Competitive cellular and extracellular interactions among endosymbionts	Exclusion of one endosymbiont by the presence of another, or facilitation if host fitness benefits	Interactions among Spiroplasma and Wolbachia in spider mites affect fitness. Higher fecundity & survival in mites coinfected	Frequency of endosymbiont cannot be predicted unless other microbes considered	Zhu et al. 2020
(xii) Presence of multiple strains of endosymbiont in the same individual or population	Direct interactions between endosymbiont strains in host or indirect effects such as through CI patterns allowing one strain to invade a population colonized by another	Likely transient unless doubly infected strains are favoured	Wolbachia associate with most insects, yet multiple variants in host individuals are relatively rare; replacement of one Wolbachia strain by another in D. simulans	Presence of multiple strains in same host and stability of transmission; replacement of strains should be common over evolutionary time but rarely captured	Weinert et al. 2015; Kriesner et al. 2013; Turelli et al. 2022

In the absence of nutritional fitness benefits, endosymbionts may spread and reach a high frequency through impacting reproduction as illustrated by extensive research on Wolbachia in multiple species (reviewed in O’Neill et al. 1997) (Table 1(ii)). A dramatic example of such an impact involves endosymbiont-induced parthenogenesis, where the endosymbiont triggers female-only reproduction and infected individuals transmit the endosymbiont to offspring (Hurst et al. 1999; Katsuma et al. 2022). Several endosymbionts cause cytoplasmic incompatibility (CI) that provides females with Wolbachia a relative fitness advantage. CI kills embryos when males carrying Wolbachia mate with females lacking Wolbachia, whereas females carrying the same Wolbachia strain are protected from CI (Hoffmann & Turelli 1997; Shropshire et al. 2020; Turelli & Hoffmann 1995). It is tempting to think of Wolbachia as “manipulating” the host in these situations to further their own spread, but this ignores positive effects on host fitness that are required for Wolbachia to initially spread to an appreciable frequency (Hoffmann & Cooper 2024; Turelli et al. 2022). This frequency-dependent nature of CI effects on hosts results in weak purifying selection on CI loci (known as CI-causing factors or cifs) (LePage et al. 2017) that tend to degrade through time (Martinez et al. 2021; Shropshire et al. 2024; Turelli 1994).

Interactions with environmental conditions like temperature can influence the strength of incompatibilities. For example, high temperatures can generate incompatibility within transinfected mosquito populations when females with a low Wolbachia density mate with transinfected males (Ross et al. 2019a), reducing the effectiveness of CI-based replacement interventions. This can reduce equilibrium frequencies in populations where CI might normally push endosymbionts and their application-relevant traits towards fixation. High temperatures up to 39 ° or 40 °C decrease the hatch rates of Wolbachia infected Dipteran females to as low as 25% when they mate with males to which they have become partly incompatible due to density changes, reducing the effectiveness of CI as a driver of an increase in Wolbachia frequency (Ross et al. 2019a).

Endosymbiont frequencies also depend on their transmission fidelity, which in turn depends on environmental conditions (e.g. temperature and nutrition) (Table 1(iii)). High temperatures reduce Wolbachia transmission in mosquitoes (Hoffmann et al. 2015) and spider mites (Van Opijnen & Breeuwer 1999), sometimes leading to Wolbachia loss (Ross et al. 2019a). The density of endosymbionts in a host can influence transmission and is often decreased by high temperatures, a phenomenon also characteristic of obligate endosymbionts such as Blochmannia in ants (Fan & Wernegreen 2013) and Buchnera in cowpea aphid (Chen et al. 2009). Wolbachia in Aedes aegypti is also lost from quiescent eggs that allows mosquitoes to persist under dry conditions (Lau et al. 2021), reducing vertical transmission rates. In the grain aphid, Sitobion avenae, transmission of Regiella endosymbionts is reduced at high temperatures (Liu et al. 2019). It seems likely that high temperatures may generally reduce endosymbiont transmission fidelity.

Low temperatures can also influence transmission fidelity, such as by reducing transmission of Wolbachia in seasonally related diapausing Nasonia vitripennis wasps (Perrot-Minnot et al. 1996) and in some Drosophila (Hague et al. 2024). At cool temperatures, wMel Wolbachia abundance is lower in the developing oocytes of Drosophila melanogaster females with a further effect of the host genetic background (Hague et al. 2022). This corresponds to temperature-dependent imperfect transmission that contributes to continent-wide latitudinal clines in wMel frequency (Hague et al. 2022; Kriesner et al. 2016). Transmission of Spiroplasma in Drosophila hydei is also decreased when they are cultured at 15/18 °C when compared to near perfect transmission at high temperatures, and other examples are found in Martins et al. (2023). Together, these data highlight that quantifying endosymbiont transmission rates in the context of the environment – including the potential for horizontal transmission discussed below (Table 1(iv)) (Gu et al. 2023) – is critical for understanding their frequencies.

The effects of facultative endosymbionts on host fitness also typically depend on abiotic and biotic components of the environment, influencing endosymbiont frequencies in specific situations (Table 1(v, vi)). The effects of wMel Wolbachia on D. melanogaster fitness varies with environmental conditions (Fry et al. 2004; Olsen et al. 2001). Females with Wolbachia have a relatively low fecundity under cold conditions, contributing to clinal latitudinal variation in Wolbachia frequencies (Kriesner et al. 2016). High temperatures impact fitness effects of other endosymbionts on hosts, as summarized in Martins et al. (2023). To understand how host fitness effects might drive endosymbiont dynamics, it is therefore important to consider field conditions (see below).

Parasitoid wasps and predators in an environment are expected to influence endosymbiont dynamics. Laboratory studies on aphids indicate that some endosymbiont strains of Regiella and Hamiltonella protect hosts against parasitoids (e.g. Scarborough et al. 2005; Vorburger 2018). An example of a predator interaction mediated by endosymbionts is production of the toxin pederin by uncharacterized endosymbionts (referred to as PPE or pederin-producing endosymbionts) in the staphylinid rove beetle, Paederus fuscipes, which provides hosts protection against predators (Fang et al. 2023). In these cases, endosymbiont frequency should increase with the presence of predators in the environment, although this is not necessarily borne out by field data as in the case of Hamiltonella (Smith et al. 2021). One issue is that an endosymbiont providing protection may have other fitness costs for a host as in the case of PPEs in rove beetles (Xiao et al. 2024). Parasitoids and predators may also influence endosymbiont dynamics by moving endosymbionts between host species (Table 1(iv) (Gehrer & Vorburger 2012; Heath et al. 1999).

Endosymbiont frequencies can also be influenced by the host genetic background (Table 1(vii)). In clonal organisms like aphids, this can result from differences in fitness effects or transmission fidelity of endosymbionts among clones. For instance, in pea aphids, the bacterial endosymbionts PASS (“Pea Aphid Secondary Symbiont”) and Rickettsia increase fecundity at 25 °C in some parental clones but not in others (Chen et al. 2000) while Regiella transinfected into different clones increased performance in one clone but not in four others (Ferrari et al. 2007). In sexual species, difference among genetic lines can also be apparent as in the case of Rickettsia affecting the fitness of whitefly lines (Cass et al. 2016). Suppression of reproductive phenotypes may evolve, as observed for Wolbachia-induced male killing in both the butterfly Hypolimnas bolina which maps to a single region (Reynolds et al. 2019) and in Drosophila pseudotakahashii where a dominant genetic factor likewise mapped to one chromosomal region (Richardson et al. 2023). Host genetic factors also contribute to Wolbachia-induced CI strength in several host systems, including mites and flies (Cooper et al. 2017; Reynolds & Hoffmann 2002; Wybouw et al. 2022). An early example involves wMel that causes weak CI in D. melanogaster (Reynolds & Hoffmann 2002), but near complete CI (no or few eggs hatch) in transinfected Ae. aegypti used in deliberate field releases. This suggests the presence of CI suppressors in the D. melanogaster genome absent in the novel Ae. aegypti host. These modifications likely involve changes in endosymbiont abundance in different tissues, which we discuss below.

For clonal organisms, the incidence of endosymbionts in clonal lineages will vary depending on the past colonization history of the endosymbionts. Co-segregation of clones and endosymbionts could change the population incidence of the endosymbionts depending on clone fitness and how this might be impacted by resident endosymbionts (Table 1(viii)). Endosymbiont-clone associations will be modified if there is horizontal transfer across clones such as through plant tissues or parasitoids. Clone-endosymbiont associations are known from the field, as in the grain aphid Sitobion avenae where one of four common clones was infected with Hamiltonella and the others varied in the incidence of Regiella (Zepeda-Paulo et al. 2017). Fitness differences among clones could therefore drive endosymbiont dynamics although this will be influenced by sexual cycles and other evolutionary processes that influence clonal composition (Loxdale 2008).

Endosymbiont density and incidence can be associated with sex (Table 1(ix)). Theory predicts that selection should favour reliable transmission of endosymbionts to female but not male offspring and potentially selection against the expression of CI by males through host genes (Tortosa et al. 2010; Turelli 1994). This could help explain why males may lack endosymbiont strains such as in CI-inducing Wolbachia in Drosophila pseudotakahashii (Richardson et al. 2019) and in Aedes albopictus typically infected by two Wolbachia strains but with males often lacking one of them (Tortosa et al. 2010; Yang et al. 2022). However, transmission to females can also be less reliable than to males at least under cool conditions, as observed for wYak in Drosophila yakuba and others (Hague et al. 2020a). It is unclear if this unexpected pattern could be explained by the relatively short association time of these Wolbachia with their current hosts. Sex-related differences in endosymbiont presence are typically mediated by changes in endosymbiont density – decreasing as males age and develop, with consequences for endosymbiont-mediated traits such as CI strength in Drosophila pseudotakahashii (Richardson et al. 2019).

Related endosymbiont strains often differ in their host phenotypic effects, such as strains of Regiella in aphids differing in the level of parasitoid protection they provide (Jamin & Vorburger 2019) and strains of Spiroplasma differing in the presence of male killing (Ballinger & Perlman 2019). In addition, endosymbionts can evolve to influence fitness traits (Table 1(x)), such as Wolbachia in D. simulans which had a negative impact on fecundity but then evolved over a 20-year period to have a positive impact (Weeks et al. 2007). Positive fitness effects are predicted even when endosymbionts cause CI for strains to initially spread (Hoffmann et al. 1990; Turelli et al. 2022).

Many endosymbiont effects on hosts are likely to be mediated by their tissue distribution. Endosymbionts like Wolbachia are widely distributed throughout host tissues, influencing transmission rates and host traits (Serbus et al. 2008). Wolbachia are often abundant in somatic tissues (Dobson et al. 1999; Pietri et al. 2016), corresponding to their effects on host behaviours that include sleep and activity, thermal preference, and pathogen protection (Albertson et al. 2013; Hague et al. 2020b; Moreira et al. 2009). Both Wolbachia (Martinez et al. 2017) and host (Emerson & Glaser 2017; Hague et al. 2022; Strunov et al. 2022) genomes contribute to their abundance and tissue distribution. Variation in density has been related to host traits but there are exceptions. For example, while the strength of CI caused by wMel in D. melanogaster declines sharply with male age, wMel abundance in testes increases with male age (Shropshire et al. 2021); and in D. simulans, wRi causes consistently strong CI as males age (Reynolds & Hoffmann 2002), but wRi abundance in testes decreases (Shropshire et al. 2021). Other factors (e.g. the expression of cif genes) may be better predictors of CI strength (Shropshire et al. 2022).

Endosymbionts can interact with other microbes when present in the same host (Table 1(xi)). Where multiple endosymbionts occur in the same host, there could be competition for resources or even for transmission; on the other hand, facilitation is also possible. Examples include a reduction in the fecundity and other fitness traits of Bemisia tabaci whiteflies that carry both Hamiltonella and Cardinium with coinfection also decreasing the density of one of the endosymbionts (Zhao et al. 2018) when compared to enhanced survival and fecundity in spider mites when Wolbachia and Spiroplasma are both present (Zhu et al. 2020). Microbiome studies show widespread associations between endosymbionts and the gut microbiome (Gupta & Nair 2020), which could in turn influence the fitness effects of endosymbionts on hosts. A complex interaction between viruses and endosymbionts could influence the incidence of both types of microbes. This extends from blockage of arbovirus transmission by Wolbachia in Aedes mosquitoes (Moreira et al. 2009) to endosymbiont-virus interactions in pea aphids where Hamiltonella increases PEMV (Pea Enation Mosaic Virus) transmission (Sanches et al. 2023), but a Regiella strain protects against APV (Acyrthosiphon pisum Virus) (Higashi et al. 2023). Likewise in oat aphids, Regiella protects against BYDV (Barley Yellow Dwarf Virus) acquisition (Yu et al. 2025). These interactions could influence endosymbiont frequencies through virus effects on host fitness and indirectly by viruses influencing plant health (Sanches et al. 2023).

When multiple strains of the same endosymbiont are present in populations or even in the same host, they could interact to influence strain frequencies (Table 1(xii)). Strain variation can be particularly marked in endosymbionts like Serratia and Arsenophonus where some strains live partly independently of hosts while others are tightly linked to hosts and these endosymbionts can show a high level of diversity even within an insect group (Jousselin et al. 2013; Renoz et al. 2021). The impact of strain variation on endosymbiont dynamics is particularly well studied in Wolbachia in Drosophila; for instance, the wRi strain in D. simulans has rapidly displaced a resident wAu strain likely due to CI (Kriesner et al. 2013). The horizontal spread of some Wolbachia strains such as wRi and wMel among divergent host species can also be relatively rapid (Cooper et al. 2019; Martinez et al. 2021; Turelli et al. 2018), while Wolbachia prophages termed Woviruses and the cifs they carry are gained and lost even more rapidly among Wolbachia genomes (Shropshire et al. 2024). Horizontal transfer of cif genes underlies dramatic variation in CI strength among closely and distantly related Wolbachia (Martinez et al. 2015; Shropshire et al. 2024), which directly affects Wolbachia frequencies.

Clearly many variables influence endosymbiont dynamics in populations and there is an increasing understanding of these effects at the cellular and tissue levels with generalizations emerging. Most endosymbiont transfers to new hosts in nature probably fail because conditions for establishment are not met – failure is difficult to detect in nature but is a hallmark of deliberate transfer attempts between aphid species (Gu et al. 2024b). Once transferred, endosymbiont frequencies will depend on effects on host fitness, reproduction and transmission which vary with the environment, and endosymbiont density often plays a critical role. Hosts can evolve to regulate and influence endosymbiont phenotypes, as can the endosymbionts. Endosymbionts may interact with other microbes present in the same individuals and populations. The above considerations make it challenging to model endosymbiont dynamics without considering it on a case-by-case level, yet this exercise is essential in thinking about how endosymbionts are used to suppress impacts of invertebrate pests and disease vectors.

3. Expanding applications

Targeting secondary endosymbionts for pest and disease control is still in its infancy but several applications involving the release of Wolbachia in Aedes mosquitoes have progressed to an operational stage. Infected males that trigger CI have been successfully released (Lim et al. 2024; Zheng et al. 2019) to suppress pest Aedes mosquito population size (Table 2(i)), while the replacement of females from natural populations lacking Wolbachia with those that carry a Wolbachia strain capable of suppressing arbovirus transmission by Aedes mosquitoes (Table 2(iv)) is now widely used. These strategies have been highly successful in some scenarios for suppressing population size and decreasing dengue incidence (Hoffmann et al. 2024; Pinto et al. 2021; Utarini et al. 2021) but have also encountered challenges. In the case of replacement, these include challenges due to environmental factors that influence fitness, transmission and the expression of CI as well as host background effects on phenotypes that include virus suppression, incompatibility and insecticide susceptibility (Garcia et al. 2019; Liang et al. 2022; Ross et al. 2019b). In the case of suppression, they include inaccurate sexing leading to the accidental release of females carrying the endosymbiont that produce offspring no longer incompatible with released males (Pagendam et al. 2020; Zheng et al. 2019).

Table 2.

Impact of endosymbiont polymorphisms on applications of endosymbiont technology.

Intervention	Approach	Examples (realized and potential)	Challenges	Example references
(i) Suppression and extinction due to CI generated by released males	Males released into a population carrying endosymbionts that generate CI when mating with field females	Use of Wolbachia strains in Aedes aegypti and Ae. albopictus to suppress mosquito populations transmitting arboviruses	Inaccurate sexing can lead to females carrying the endosymbiont being released, disrupting CI from males, also requires large release effort, community concerns	Lim et al. 2024; Pagendam et al. 2020; Zheng et al. 2019
(ii) Allow CI within lines to decrease population size	Likely a side effect of an intervention aimed at CI-based suppression or replacement where polymorphism develops	Persistent polymorphism has developed in mosquito releases with Wolbachia CI but population effects have not been measured	Requires persistent polymorphism, only partial suppression of population	Not specifically tested but see: Ross et al. 2019b; Ahmad et al. 2021; Pinto et al. 2021
(iii) Releases with strains carrying endosymbionts decreasing fitness	Release pest strains with endosymbiont repeatedly to decrease population size at critical periods	Suggested and modelled for Rickettsiella in aphids, also tested experimentally for Wolbachia with deleterious effects on Aedes mosquitoes at a critical egg stage	Requires a mode of transmission of endosymbiont outside of host such as plant-based transmission or spread via CI	Gu et al. 2023; Slavenko et al. 2024; McMeniman & O’Neill 2010; Ross et al. 2020; Ritchie et al. 2015
(iv) Replacement of host population with endosymbionts that reduce microbe-causing disease transmission	Releases for replacement require CI or another mechanism to spread the endosymbiont and help it maintain high frequencies	Wolbachia in Aedes to decrease dengue incidence, Wolbachia in planthoppers and Regiella in aphids to decrease plant virus transmission	Spread can be challenging when environmental conditions influence fitness of host carrying endosymbiont and vertical transmission, community concerns can be an issue	Hoffmann et al. 2011; 2024; Pinto et al. 2021; Ahmad et al. 2021; Utarini et al 2021; Sanches et al. 2023; Higashi et al. 2023; Yu et al. 2025
(v) Replacement of host population for desirable features other than microbe transmission interference, (e.g. chemical resistance)	Since endosymbionts affect a range of host traits including insecticide susceptibility and habitat choice, replacement could be combined with other interventions to suppress pest populations	No directed applications but a range of microbes alone or in combination e.g Serratia, Arsenophonus and Regiella could influence useful traits	As for (iv)	Not yet tested
(vi) Chemical applications targeted to remove essential endosymbionts that increase host fitness	Suppression of primary endosymbiont could decrease pest fitness, as could removal/suppression of secondary endosymbionts providing nutrition or natural enemy protection	Chemicals (e.g. fungicides) can have antibiotic effects affecting endosymbionts although removal of primary endosymbiont in aphids has proven difficult	Chemicals can have off target effects, need to be integrated into overall IPM strategy	Gao et al. 2021; 2023; Chirgwin et al. 2022
(vii) Introduction of self-spreading endosymbionts > reproductive modifications like male killing, affecting populations when males become rare	Needs male-killing endosymbionts capable of increasing such as through host benefits	Population suppression/ extinction has been documented in Lepidoptera but host evolution to counter effects has been observed in Drosophila and Lepidoptera	Could be countered by strong host selection affecting killing phenotype, concerns similar to those for replacement releases	Hornett et al. 2009; Richardson et al. 2023
(viii) Improved host fitness by natural endosymbiont or transinfections generating parthenogenesis	Releases with enemies such as wasp parasitoids that are parthenogenetic	Biological control becomes more effective when using parthenogenetic strains (no male wastage)	Challenges in generating locally adapted strains when parthenogenesis is present	Russell et al. 2016
(ix) Improved host fitness in natural enemies due to effects on fitness-related traits	Releases with enemies that show other useful traits that increase their effectiveness	Biological control becomes more effective due to traits like host dispersal as observed for Rickettsia in Erigone atra)	Fitness costs associated with endosymbionts producing useful traits	Goodacre et al. 2009
(x) Altered plant defence due to presence/absence of endosymbionts	Presence of endosymbionts could increase virulence of pests on weeds and decrease impact on crops	Endosymbiont diversity in spider mites related to changes in defence of tomato plants	Needs caution because interaction with crops could be negative such as for Russian wheat aphid and Rickettsiella	Ribeiro et al. 2020

These interventions in Aedes mosquitoes have used endosymbionts transinfected across species which are deliberately introduced into target insects and generate larger effects on hosts (in terms of CI, immune regulation, fitness etc.) than native endosymbionts (Ross et al. 2019b). For instance, while the wMel Wolbachia of D. melanogaster may provide weak protection against viruses (Cogni et al. 2021) – which are hard to detect (Shi et al. 2018) and virus-specific (Cogni et al. 2021) – the impact of transinfected wMel tested against dengue and other arboviruses in Ae. aegypti is strong (Moreira et al. 2009). As noted above, the experimental movement of endosymbionts between divergent donor and recipient hosts decouples existing evolved interactions in ways that may reveal endosymbiont encoded effects not observed in their natural host. In the case of stronger wMel-induced CI in transinfected Aedes, this pattern supports suppression of wMel CI by its natural host D. melanogaster. These genetic interactions may in turn depend on environmental conditions, highlighting the need for field-based analysis of host-endosymbiont interactions and estimation of critical parameters that govern endosymbiont spread in natural settings.

In contrast, most experiments on agricultural pest species have focused on natural associations or transinfections within species (Gu et al. 2024b). Successful transinfections of Wolbachia have taken place in fruit flies and planthoppers (Gong et al. 2023; Gong et al. 2020), but some pest species including aphids appear resistant to Wolbachia transinfections (Gu et al. 2024b). Nevertheless, because CI caused by Wolbachia and other symbionts including Cardinium (Mann et al. 2017) is widespread in insects (Turelli et al. 2022), it should be possible to use endosymbionts to undertake releases of males to suppress pest populations in much the same way as sterile releases are undertaken against agricultural pests like Mediterranean fruit flies (Enkerlin 2021). There is a benefit in using Wolbachia because the high doses of radiation typically used in male sterile releases are not required to induce female sterility, reducing radiation-related fitness costs in males. However, accurate sexing will still be required (Gong et al. 2024), particularly in the absence of radiation, to prevent the establishment of female lineages resistant to CI induced by the release strain (Pagendam et al. 2020). In Aedes mosquitoes, this concern has led to programs that combine low rates of irradiation to sterilise females that have escaped the sexing process with releases of Wolbachia infected males that have also been irradiated but remain competitive (Zheng et al. 2019). Reliable large-scale sexing has not yet been developed for agricultural pests carrying transinfections (Gong et al. 2023).

Endosymbionts could be introduced to generate CI and partial suppression in populations (Table 2(ii)). Incompatible matings between males carrying endosymbionts generating CI and females lacking them could decrease population fitness. With an endosymbiont at a frequency of 50%, 25% of matings may be incompatible and decrease population size. This may seem like a minor effect but impacts of pests typically depend on them exceeding economic thresholds required for crop damage. This intervention requires persistent polymorphism which is typical of many natural Wolbachia associations (Hoffmann and Cooper 2024) and could be induced through deliberate releases. Host costs associated with an introduced endosymbiont could result in further population suppression in multivoltine species. An example is the use of the wMelPop Wolbachia transinfection in Aedes aegypti which causes a sharp reduction in egg hatch when eggs are in a quiescent state in a dry season (McMeniman & O’Neill 2010; Ross et al. 2020). A combination of CI and the expression of this fitness effect could suppress or even eliminate a population once CI spreads the Wolbachia during a favourable wet season, as validated in large cage experiments with Aedes aegypti carrying the wMelPop infection (Ritchie et al. 2015).

For clonal species, it is not possible to take advantage of CI when introducing endosymbionts but novel transinfections can suppress populations in other ways (Table 2(iii)). Pathogenic endosymbionts like Rickettsiella can be readily transinfected across aphid species where they may have large deleterious effects. Persistence of this endosymbiont depends on horizontal transmission through plant tissue and high vertical transmission fidelity (Gu et al. 2023) (Table 1(iv)(Fig. 1). Models indicate that populations of the green peach aphid (Myzus persicae) could be suppressed below economic thresholds by repeated releases of aphids with Rickettsiella (Slavenko et al. 2024).

Fig. 1.

Potential impact of Rickettsiella transinfection in Myzus persicae on pest aphid suppression following horizontal transmission coupled with deleterious fitness effects. Based on data and images from Gu et al. (2023) and on simulation images of population-level impact from Slavenko et al. (2024).

A replacement or suppression strategy could be applicable to several agricultural pests such as leafhoppers, planthoppers, whiteflies and aphids that spread plant diseases (Table 2(iv)). A Wolbachia strain (wStri) transinfected from one plant hopper species (Laodelphax striatellus) to another one (Nilaparvata lugens) effectively suppressed transmission of rice ragged stunt virus and could spread via CI. Although virus transmission may be enhanced by native endosymbionts (Sanches et al. 2023), some strains of Regiella can suppress viruses of hosts (Higashi et al. 2023) and plants including the transmission of barley yellow dwarf virus by transinfected Rhopalosiphum padi (Yu et al. 2024).

Beyond disease transmission, replacement strategies could introduce other desirable traits associated with endosymbionts that affect pest biology to facilitate their control (Table 2(v)). These include reduced movement rates (Gu et al. 2024a), or increased susceptibility to agrichemicals in pest species (e.g. Pang et al. 2018). Endosymbionts can decrease as well as increase chemical susceptibility (Liu & Guo 2019) or have little impact (Leybourne et al. 2023) so traits need to be characterized within the context of specific endosymbiont strains and species/genetic backgrounds, particularly given that transinfections can have quite different impacts than native associations even when the same endosymbiont strain is involved. Native endosymbionts like Hamiltonella may decrease aphid susceptibility to pesticides and natural enemies (Li et al. 2021). However, different strains of the same bacterium can show quite different phenotypes as in the case of the impact of Arsenophonus strains on pesticide susceptibility (Pang et al. 2018). Moreover, gut bacteria may influence insecticide resistance through their detoxification capacity (Kikuchi et al. 2012) so the interaction between these microbes and endosymbionts needs to be considered. Few studies have considered the impact of interspecific transinfections on pesticide resistance, but transinfected Rickettsiella in green peach aphids (Dorai et al. 2024) and Wolbachia in Aedes mosquitoes (Endersby & Hoffmann 2013) have little impact on susceptibility to a range of chemicals.

Where endosymbionts are important for host fitness, it may be possible to apply some agrichemicals or other stressors to decrease the density of endosymbionts and thereby decrease host fitness (Table 2(vi)). This follows from the antibiotic properties of some agrichemicals including fungicides and herbicides. Strobilurin fungicides both influence the fitness of cotton aphids and the density of the primary Buchnera endosymbiont (Gao et al. 2023), while the herbicide glyphosate disrupts the microbiome (though not specifically endosymbionts) of Harmonia ladybirds (Gao et al. 2021). A challenge is to establish direct connections between fitness effects of chemicals and impacts on primary endosymbionts (Chirgwin et al. 2022).

The idea that some endosymbionts could cause reductions in pest population size by changing host reproduction to increase their own transmission (Table 2(vii)) has emerged from studies of natural insect populations. Repeat collections of the tropical butterfly Hypolimnas bolina showed wildly fluctuating sex ratios in populations associated with a male killing Wolbachia strain (Hornett et al. 2009). While host genes also eventually evolved to suppress the endosymbiont effect on sex ratio in this species, perhaps some hosts cannot evolve suppression and these populations might then be suppressed. Male killers involving Spiroplasma as well as Wolbachia are well known in Diptera, Lepidoptera and some other insect orders (Ballinger & Perlman 2019) but rarely explored in pest species.

A potentially important but rarely explored use of endosymbionts is to generate parthenogenetic lines in natural enemies showing haplodiploidy (Table 2(viii)). Parasitoid wasps are widely released as biocontrol agents at critical times. Wasps including the egg parasitoid Trichogramma and many other species can be parthenogenic due to the presence of Wolbachia (Huigens & Stouthamer 2003) and other endosymbionts including Rickettsia (e.g. Giorgini et al. 2010). The commercial production of sexual lines of these parasitoids is challenging due to inconsistent and often male-biased sex ratios, which reduces the efficiency of releases given that only females parasitise pests. Parthenogenetic lines could reduce production costs and increase field efficacy particularly when released wasps can immediately parasitise pests without mating (Liu et al. 2018). Whether parthenogenetic strains generated through endosymbiont introductions are useful will depend on the penetrance of the parthenogenetic phenotype, whether parthenogenetic females produce offspring without mating, and fitness costs. Female Trichogramma and other parasitoids infected by some Wolbachia strains may show fitness costs for some life history traits (Zhou et al. 2021) but costs tend to vary among strains (Russell et al. 2016). Careful optimisation of production systems for parthenogenetic strains (Zhou et al. 2020) and the use of novel approaches to overcome deleterious effects (Lindsey & Stouthamer 2017) could help overcome these issues.

Beyond being used to generate parthenogenetic strains, endosymbionts could also improve biocontrol agents in other ways (Table 2(ix)). In the spider, Erigone atra, which is important for pest control, Rickettsia endosymbionts are associated with increased dispersal (Goodacre et al. 2009), an important attribute of successful biocontrol. Endosymbionts with useful attributes could be introduced into beneficial species through transinfection. Conversely, native endosymbionts could be removed if they have costs. For instance, Wolbachia generates life history costs in the moth Ephestia kuehniella (Sumida et al. 2017) which is often used as a factitious host to rear parasitoids; here removal of the endosymbiont could improve strain quality.

Finally, the way endosymbionts interact with plants has rarely been explored but could provide new applications for pest control (Table 2(x)). Transinfections of Russian wheat aphid by Rickettsiella enhance plant virulence resulting in more rapid plant death in wheat, barley and grassy weeds (Gu et al. 2024a). While this could increase crop damage where infestations take place, it could also help control weeds comprising the green bridge for Russian wheat aphid across non-cropping periods. Natural Serratia endosymbionts can suppress plant defences in pea aphids feeding on Medicago (Wang et al. 2020) and microbiome diversity in spider mites may be related to plant defence suppression in tomatoes (Ribeiro et al. 2020). It remains to be seen if parasitic endosymbionts (particularly as transinfections) enhance plant defences and interact more generally with the plant microbiome.

In considering these applications, there will always be concern that evolutionary changes can overcome the benefits provided by endosymbionts. The rate at which host suppression of CI may emerge is difficult to predict, although natural D. melanogaster suppression of CI must have evolved since it acquired wMel in the last few 100 thousand years (Shropshire et al. 2024). Degradation of cifs themselves due to weak purifying selection occurs on the order of tens to hundreds of thousands of years (Beckmann et al. 2021; Meany et al. 2019), leading terminally to rescue-only phenotypes as predicted by theory (Turelli et al. 1994). It seems unlikely that host suppression of CI or cif degradation will occur rapidly enough to influence the efficacy of Wolbachia applications. However viral evolution or the evolution of host-encoded suppression of endosymbiont effects will need to be considered.

4. Field data

Much of the research on endosymbiont effects on hosts and other microbes is laboratory based. The number of factors impacting endosymbiont frequencies and densities is clearly large so that the field dynamics of endosymbionts – and deliberately released endosymbionts in particular – is not straightforward. Native secondary endosymbionts typically show a remarkable range of polymorphism across populations and across time that provide challenges in predicting in-field behaviours. Here are some examples.

1. Spider mites show a variable incidence of Wolbachia, Cardinium, Spiroplasma and Rickettsia in populations and across related species. Moreover, there are often interactions among endosymbionts as reflected by a high incidence of coinfection by Wolbachia and Cardinium in China (Zhang et al. 2016) and in Europe (Zélé et al. 2018).

2. Native Wolbachia associations in Drosophila often vary in frequency across space and time. Wolbachia frequencies in species may be near 100% as in D. pandora (Richardson et al. 2019) or at low-intermediate levels such as in D. innubila (Dyer & Jaenike 2005), the D. yakuba clade (Cooper et al. 2017; Hague et al. 2020a) and D. mauritiana (Giordano et al. 1995; Meany et al. 2019). Population patterns are affected by the reproductive effects of the strains (e.g. CI versus male killing) but also by imperfect transmission and fitness effects that may depend on environmental factors which produce latitudinal patterns in frequencies such as in D. melanogaster (Hague et al. 2022; Kriesner et al. 2016). Changes in endosymbiont frequencies across time have been documented in several Drosophila species including the rapid spread of the wRi Wolbachia associated with CI in Californian D. simulans (Turelli & Hoffmann 1991), and in weak-CI causing D. yakuba clade species where the frequencies of Wolbachia closely related to wMel vary across years – and between islands – in west Africa (Cooper et al. 2017). Wolbachia frequencies may also vary on seasonal timescales, as observed for Wolbachia closely related to wRi that associate with the D. suzukii pest hosts (McPherson et al. 2023).

3. In Sitobion avenae aphid hosts, changes in the incidence of Regiella, Serratia and Hamiltonella endosymbionts across a season have been documented, with Serratia and Hamiltonella predominating in December to February (winter) but Regiella dominating in March to May (spring) (Liu et al. 2019). A complication in aphid studies is that endosymbiont changes can depend on clonal aphid genotypes wherein they are common, as in the case of Regiella in Sitobion (Zepeda-Paulo et al. 2017), with clonal frequencies also being influenced by the environment.

In addition to these examples involving native endosymbionts, deliberate releases of transinfected Wolbachia in field Aedes mosquitoes also indicate variability in outcome. The initial releases of Wolbachia into uninfected Aedes aegypti in Australia resulted in a rapid increase in Wolbachia frequency to near fixation through strong CI and likely minor fitness costs (Hoffmann et al. 2011), a pattern duplicated in other Australian locations (Ryan et al. 2019) and in Yogyakarta in Indonesia (Utarini et al. 2021). On the other hand, Wolbachia releases in Kuala Lumpur, Rio and other locations have not necessarily led to high and stable equilibrium frequencies (e.g. Ahmad et al. 2021; Pinto et al. 2021). Instead, Wolbachia frequencies have often increased in some local areas but not others. Several factors related to the local ecology, genetic background and strain attributes are likely to contribute to these diverse outcomes in nature, although the precise reasons involved often remain undefined. Leveraging knowledge of strain divergence and patterns of local adaptation are obvious ways forward.

5. Concluding statements

Endosymbionts could have many potential applications in pest and disease control particularly through transinfections, but most of these are not yet realised. As a starting point, it is important to understand the diversity and distribution of native endosymbionts in target pests and natural enemies which may influence a range of traits such as susceptibility to natural enemies, chemicals, virulence and the capacity to transmit plant viruses. This is then followed by applications particularly based on transinfections which require an understanding of the diversity of strains of endosymbionts within a species and across related species, often with host-specific effects and different types of interactions among microbes and environments. The extensive experience with Wolbachia in mosquitoes indicates how endosymbionts can be successfully used in applied settings. Nevertheless, applications with other novel endosymbionts require additional work, particularly in natural settings.

Moreover, safety issues will need to be considered for all interventions. Several risk assessments have been undertaken prior to deliberate releases of mosquitoes involving transinfected Wolbachia. The first risk assessment was undertaken in Australia by the CSIRO (Murray et al. 2016) and has been followed by assessments in other countries such as in Indonesia (Buchori et al. 2022). The broad conclusion from these studies is that risks associated with Wolbachia releases are extremely small, further borne out by more than 13 years of releases without documented adverse outcomes during many successful field demonstrations of disease interventions (e.g. Utarini et al. 2021). However, safety issues will need to be considered on a case-by-case basis, building on extensive work undertaken in Wolbachia, with some general issues considered in the Supplementary File. In summary, decades of research on natural endosymbionts could now be leveraged to identify specific applications, particularly through the use of transinfections, whose impacts need to be evaluated in the context of environmental and genetic interactions to facilitate their successful and safe application.