Reproductive Dominance Strategies in Insect Social Parasites
- 314 Downloads
In eusocial insects, the high cost of altruistic cooperation between colony members has favoured the evolution of cheaters that exploit social services of other species. In the most extreme forms of insect social parasitism, which has evolved multiple times across most social lineages, obligately parasitic species invade the nests of social species and manipulate the workforce of their hosts to rear their own reproductive offspring. As alien species that have lost their own sociality, these social parasites still face social challenges to infiltrate and control their hosts, thus providing independent replicates for understanding the mechanisms essential to social dominance. This review compares socially parasitic insect lineages to find general trends and build a hypothetical framework for the means by which social parasites achieve reproductive dominance. It highlights how host social organization and social parasite life history traits may impact the way they achieve reproductive supremacy, including the potential role of chemical cues. The review discusses the coevolutionary dynamics between host and parasite during this process. Altogether, this review emphasizes the value of social parasites for understanding social evolution and the need for future research in this area.
KeywordsInquilinism Fertility signal Queen pheromone Coevolution Eusociality
Parental care can be a substantial investment, especially for social species that invest across generations. Social investment can reap strong benefits, as can be seen by the success of social hymenopterans (ants, bees and wasps) in terms of species diversity, geographic range, biomass (Fittkau and Klinge 1973; Wilson 1987), and phylogenetic longevity (Barden and Grimaldi 2016). Such systems, however, are particularly vulnerable to cheaters (predators, parasites, and social parasites) that seek to exploit the wealth of resources social species amass. Such individuals and species that capitalize on the investments of other species exemplify the disadvantages of sociality, and thus are instrumental in the evolution of intricate and innovative social defenses. Understanding the mechanisms established cheaters utilize in these systems can help us gain valuable understanding into the mechanisms vital to social interactions and help us tease apart the process of social evolution.
Most eusocial insect colonies consist of only one reproductive individual, the queen, and non-reproducing workers (Wilson 1971). Given the disadvantages of forfeiting reproduction, these systems must establish robust mechanisms to maintain dominance of the fertile individual in the nest while keeping strong cooperation and mutual assistance between all colony members. This review brings attention to the value of social parasites for understanding the mechanisms regulating reproductive dominance in social insects. First, we highlight the diversity of social parasites and emphasize the types of social parasites that could help with filling this particular gap in our understanding of insect sociality. We then present the different challenges that are faced by social parasites for successful usurpation of host nests. We focus in particular on the hurdle of achieving reproductive supremacy in host colonies and how host social organization and social parasite traits impact the behavioral and chemical strategies used to overcome this hurdle.
The Breadth of Social Parasitism
Social parasitism in insects can be broadly defined as the interaction between two species in which a parasite gains benefits at the expense of a social host (Hölldobler and Wilson 1990; Schmid-Hempel 1998). These benefits can be of many sorts and vary in the scope of their parasitism. To the least detriment, some social parasites utilize colonies as shelters or protection against predators. A more negative impact is exhibited by social brood parasites that exploit stores or host brood as a food resource. In this review, however, we refer to the most extreme form of social parasitism – social parasitism sensu stricto – whereby parasites seek not merely to utilize the resources of the nest, but to utilize and control the workforce within it to feed and care for their own reproductive offspring (Buschinger 1986).
Social parasites s. str. are derived from eusocial ancestors and retain many social traits of their related eusocial hosts. They are especially common within hymenopterans, with instances in the ants, social wasps, and social bees, where a large spectrum of alternative modes of social parasitism have evolved. Among these different types of parasitic interactions, socially parasitic strategies exhibit a gradation from facultative to obligate. Facultative socially parasitic species can nest independently (Nash and Boomsma 2008) but tend to usurp nests of their own or other closely-related species (Alford 1975). This is generally considered a first step in the evolution of specialized social parasitism (Brockmann 1993; Taylor 1939; Wilson 1971) and is exhibited by many species of ants, wasps and bees.
In contrast to facultative social parasites, obligate social parasites are totally dependent on their hosts to complete their life cycle. In the most extreme form of obligatory social parasitism (inquilinism), after infiltrating host nests, the parasitic females lay eggs that only produce sexuals. They thus have lost their ability to produce a worker caste and rely on the host workers to rear their offspring (Alford 1975; Buschinger 2009; Cervo 2006). Each of the major eusocial hymenopteran lineages exhibits cases of this extreme form of social parasitism. Obligately socially parasitic species have evolved numerous times in ants with more than 200 parasitic species discovered (Hölldobler and Wilson 1990). Many of them have evolved as cheaters within a species, and thus parasitize the sister species from which they are ancestrally derived, a phenomenon known as Emery’s rule (Emery 1909). The ant genera Myrmica and Leptothorax, for example, contain multiple sister lineages comprised of a social parasite and host (Brandt et al. 2007; Savolainen and Vepsalainen 2003). In social wasps, there are 4 independent origins of obligate social parasitism: a lineage in Polistes paper wasps containing 3 species, a lineage in Dolichovespula yellowjackets with 3 species, and 2 social parasite species/lineages in Vespula hornets (Carpenter and Perera 2006; Lopez-Osorio et al. 2015). Obligate social parasitism has three origins in the bumble bees: two species attack close relatives (Brasero et al. 2018) and one diverse lineage is comprised solely of obligate social parasites (28 species) (Hines and Cameron 2010).
Obligately socially parasitic species have lost many social traits. This can include the ability to initiate their own nest, produce workers, and forage and feed offspring (Alford 1975; Buschinger 2009; Cervo 2006). At the same time these parasites have evolved new parasitic traits that enable them to capitalize on the social strategies of these systems to detect, deceive and manipulate their host (Cini et al. 2015; Smith et al. 2015b). For this purpose, they have retained some social traits, like the ability to communicate and control reproduction within the colony. Some of them are queen-tolerant in that they cohabit with the host queen who either maintains reproduction and dominance or is submissive to the social parasite and may or may not reproduce further. Others are queen-intolerant and eliminate the host queen, in this case the reproductive success of the host colony is eliminated (Brandt et al. 2005). Another extreme form of obligatory social parasitism – slavery or dulosis – is observed in ants (d’Ettorre and Heinze 2001). In this type of parasitic interaction, the slave-making species exploits the labour of host workers by regularly conducting group raids against neighbouring host colonies, capturing and enslaving the adult workers and stealing their larvae and pupae to reinforce the workforce in their nest (Mori et al. 2001). Hereafter, we limit our discussion to obligatory forms of social parasitism where a socially parasitic female takes over reproductive dominance in the host nest of another species by either replacing the queen or cohabiting with and exploiting the queen. These socially parasitic females have managed to deceive social species that are reinforced to protect themselves from cheaters and thus must exploit queen dominance strategies to attain reproductive control.
Usurpation Challenges for Social Parasites
Once the host colony is located, social parasite females face a second challenge in needing to evade host front-line defenses to successfully infiltrate the nest. To do so, most social parasites have evolved morphological features enabling them to survive host worker attacks. They are usually heavily armored to win during aggressive interactions, capable of outmaneuvering dozens of aggressive workers. In bumble bees and wasps, social parasites have a thicker cuticle and longer and/or more robust mandibles and stings that allows them to effectively resist the attacks of their hosts (Cervo 1994; Cini et al. 2011b; Fisher and Sampson 1992; Ondricek-Fallscheer 1992).
Social parasite front-line, and sometimes continuing, defenses can also include the production of defensive chemical allomones that manipulate host behavior to the advantage of the parasite. These chemical signals can include appeasement substances that reduce worker aggressiveness (e.g., decyl butyrate secreted by the Dufour’s gland of Polyergus rufescens; Mori et al. 2000) or deterrent substances, which prevent attack (e.g., possible alkadienes secreted through the cuticule of Formicoxenus nitidulus; Martin et al. 2007). These allomones can also have a repellent effect, leading to parasite avoidance, which protects parasite females from host aggressive behavior. Repellent compounds have been discovered in bumble bees (e.g., dodecyl acetate secreted by the Dufour’s gland of Bombus norvegicus; Zimma et al. 2003; tetradecyl acetate secreted by the cephalic labial glands of Bombus vestalis; Lhomme et al. 2012, 2015) and ants (e.g., decyl butanoate secreted by the Dufour’s gland of Polyergus rufescens; d’Ettorre et al. 2000; tetradecanal secreted by the Dufour’s gland of Rossomyrmex minuchae; Ruano et al. 2005). Another chemical infiltration strategy, exhibited by some slave-making ants and Polistes wasps, is the production of “propaganda” substances that induce social disorder through panic and aggression among host workers (e.g., unidentified compound secreted by the Dufour’s gland of Leptothorax kutteri; Allies et al. 1986; most likely amides secreted by the venom glands of Polistes sulcifer; Bruschini and Cervo 2011).
To bypass the host recognition code and thus reduce the need for defensive strategies, some social parasites have evolved chemical integration strategies to successfully infiltrate and maintain acceptance within the host social system. Some of them use a chemical insignificance strategy involving the failure to produce detectable cuticular hydrocarbons, a strategy utilized in some ant social parasites (e.g., very low production of C29–C35 CHCs in Acromyrmex echinatior; Lambardi et al. 2007) but also in socially parasitic wasps (e.g., CHCs produced by invading P. atrimandibularis represent only 22% of the amounts produced by its host; Lorenzi and Bagnères 2002). A second strategy, often combined with chemical insignificance, is chemical camouflage, where parasites acquire host CHC profiles through direct contact with host nest material and colony members. This strategy has been suggested in species such as the slave-making ants Polyergus rufescens (switches from CHC alkanes to host specific methyl and di-methyl branched alkanes after integration; Johnson et al. 2001), and Polyergus samurai (exhibits increased amounts of host specific alkenes and branched-alkanes in CHC profile after integration; Tsuneoka and Akino 2012) and the wasp social parasites Polistes atrimandibularis (parasite specific alkene are replaced by host specific alkanes; Bagnères et al. 1996) and Polistes sulcifer (switch from parasite specific to host specific di-methyl branched alkanes; Turillazzi et al. 2000). A final strategy is host deception through innate chemical mimicry, whereby parasites actively produce specific CHCs matching host signatures. This innate mimicry is considered more common in parasites that specialize on a narrow range of host species and has been suggested in some socially parasitic bumble bees (e.g., mimicry of host specific alkene isomere profiles in the social parasites Bombus vestalis and B. rupestris; Martin et al. 2010).
Once fully accepted inside the host colony, parasitic females face a final challenge of establishing reproductive dominance, first over the host queen if she is not killed during usurpation, but mostly over the host workers that will need to be deceived and manipulated into rearing the parasitic brood. Parasite female reproduction relies on gaining and maintaining control of the worker force, thus social parasites must maintain the ability to exploit the brood care behavior and social system of their hosts. This last hurdle allows us to dissect the essential components of reproductive dominance signaling in social insects and determine how these may vary with different types of social systems.
Factors Impacting Social Parasite Dominance Strategies
Host Social Organization
Social hymenopterans have evolved a large spectrum of societies ranging from very small colonies with nestmates having almost equal reproductive potential to huge colonies with morphologically and physiologically specialized castes and a highly skewed reproductive divison of labor (Peeters and Liebig 2009). Given the limitations of dividing eusociality into discrete categories, we characterize insect eusociality as a spectrum of social organization from simple to complex.
Colonies with simple eusociality, like some ponerine ants, some polistine wasps (paper wasps) and bumble bees, are usually small. Mature nests contain on average 20–75 individuals in paper wasps, 20–400 individuals in bumble bees and 100–1000 individuals in ponerine ants. The use of aggressive behaviors is argued to be a primary inhibitory mechanism employed by the queen or the dominant individual to prevent reproduction of the subordinate workers in these social systems (Kocher and Grozinger 2011; Ratnieks et al. 2005). Although chemical signaling may be involved, the dominance hierarchy in simple eusocial insects tends to be initiated via behavioral interactions, as suggested in some paper wasps (Pardi 1948) and in the bumble bee Bombus terrestris (van Doorn and Heringa 1986). Because all adult female members of these colonies are capable of producing offspring (at least males), behavioral dominance to suppress and control workers and their reproduction is an ongoing process necessary across the colony life cycle (e.g., in Polistes dominula; Dapporto et al. 2010; Liebig et al. 2005; Monnin et al. 2009).
In more complex hymenopteran eusocial species like most ants, some vespine wasps (yellow jackets and hornets), and honey bees, colonies may contain thousands of individuals. In these colonies, worker inhibition by the queen seems to be achieved through pheromones rather than by direct physical aggression (Le Conte and Hefetz 2008). For example, queens of the ant Lasius niger produce a cuticular hydrocarbon pheromone that inhibits worker ovary activation (3-methylhentriacontane; Holman et al. 2010). In the large colonies of complex eusocial insects, chemical signaling is more effective and often necessary to ensure adequate regulation of social interactions and reproductive division of labour (Peso et al. 2015) as it is impossible for the queen to regularly interact physically with all workers and use physical intimidation to inhibit their reproduction (Endler et al. 2004; Nascimento et al. 2004). Queens in complex eusocial species are often the only reproductives, with the workers activating their ovaries only rarely. Workers are often specialized on specific tasks regulated at the colony level (Camazine 2003), whereas in simple eusocial insect societies the control of colony activity is much more centralized and largely maintained by the queen (Green et al. 2016). Exception to these patterns do exist, however, for example Polistes dominulus has a more simple eusociality but utilizes worker as opposed to queen initiation of colony activity (Jha et al. 2006).
The nature and diversity of chemical signals used for reproductive control will also likely vary with social complexity (reviewed in Oi et al. 2015; Oliveira et al. 2015; Smith and Liebig 2017). Smith and Liebig (2017) proposed that in smaller colonies with simple eusociality, CHC profiles used for dominance-recognition are likely to be more diverse and evolutionary labile, as these complex blends are associated with dominance through learned physical interactions among individuals. In somewhat larger colonies with reduced capacity for individual recognition, a more narrow range of CHCs associated with fertility (“fertility signals”) would instead be used to recognize dominant individuals. In the largest colonies, which can no longer rely on worker contact, inhibition of worker reproduction is more likely to be the result of production of a few conserved and innately perceived “queen pheromones”. In support of this model, complex eusocial insects including Lasius ants and the vespine wasp Vespula vulgaris, characterized by large colony sizes and a strong reproductive division of labor, show inhibited ovary development when exposed to queen pheromones and, despite independent social origins, 3-methyl alkanes have been implicated as a shared major player in this signaling across social lineages (Holman et al. 2016; Oi et al. 2016; Smith and Liebig 2017). Whereas in more simple eusocial species, characterized by smaller colony sizes and overt agressive behaviors between nestmates, queens seem to exploit multiple classes of compounds and exhibit more complex and diverse fertility or dominance linked CHCs that can differ between species (e.g, in Odontomachus ants; Berthelot et al. 2017; Smith et al. 2016), or even within species, as shown in Polistes dominula (i.e., queens show inter-population variation in their dominance induced CHC composition; Dapporto et al. 2004).
Social parasite worker control strategies are likely to mirror that of their hosts and thus be impacted by the levels of host social organization. In small host colonies of simple eusocial insects, where the host queen often regulates the activity in the nest, social parasites should be forced to assume the host queens position by controlling host workers individually through direct physical contact or chemical manipulation (Green et al. 2016). On the other hand, in large colonies, where colony activity is not regulated by individual aggressive strategies, social parasites may exhibit less individual control over colony members. In such case we might expect that parasitic females would need to mimic the host queen pheromone signaling to be accepted as the dominant reproductive.
Usurpers of simple eusocial insect colonies in Polistes wasps and bumble bees are thought to take over reproductive control and inhibit worker reproduction using both behavioral dominance and chemical cues (Cervo 2006; Fisher 1983; Kreuter et al. 2012; Vergara et al. 2003). In socially parasitic bumble bees, chemical signals are thought to be transmitted to workers during physical interactions such as frequently observed head-rubbing (Fisher 1983; Küpper and Schwammberger 1995; Van Honk et al. 1981). Kreuter et al. (2012) showed that the generalist social parasite bumble bee B. bohemicus was able to suppress ovarian activation of B. terrestris host workers but only when in direct physical contact with them. The same authors also noted a significant increase in the relative amounts of the wax-type esters tetracosyl oleate and hexacosyl oleate in the B. bohemicus cuticular profile after nest usurpation, suggesting that these non-hydrocarbon cuticular compounds might play a role in host worker inhibition.
In the bumble bees, social parasite fertility signals are likely to match those of the host queen. Kreuter et al. (2012) found that the relative proportion of the CHCs of the parasitic bumble bee Bombus bohemicus is a closer match to that of the host queen than that of the workers, even in the absence of the dominant host queen. In Polistes wasps, Dapporto et al. (2004) reported similar results with the parasitic female Polistes sulcifer, which closely mimics the CHC proportion of the foundress rather than the worker profile of its host Polistes dominula. Parasite chemical mimicry of host queen signals, however, could be derived from a shared biosynthetic process related to increased ovarian activity rather than specific production for signaling purposes. These chemical signals could also be acquired in some cases through physical contacts with the host queen (Tsuneoka and Akino 2012). Further studies are needed to clarify whether worker control in more simple social species involves shared cross-species fertility signals that are retained in social parasites, whether social parasites have evolved fertility signals specific to its hosts, or if common signals between parasites and host queens are merely a by-product of ovarian development.
Although most social parasites of simple eusocial hosts seem able to inhibit worker ovarian development, it is not always the case. In the five studied social parasite bumble bees, four were able to fully suppress worker reproduction (Fisher 1983, 1984; Kreuter et al. 2012; Vergara et al. 2003) and in the three socially parasitic Polistes investigated, two were able to do so (Cervo and Lorenzi 1996; Greene et al. 1978; Jeanne 1977). For those species lacking the ability to fully control worker reproduction, an alternative might be to use an “explosive” reproduction strategy involving laying a maximum of eggs in the first days of usurpation before host workers exhibit active ovaries and start to compete for reproduction. This might be the case of the social parasite Polistes sulcifer who is only able to suppress host worker reproduction during the first weeks post-usurpation (Cini et al. 2014). Moreover, females of socially parasitic bumble bees (subgenus Psithyrus) possess more ovarioles per ovary than other bumble bees, with 6 to 18 ovarioles per ovary in social parasites and 4 ovarioles per ovary in the rest of the bumble bees (Fisher and Sampson 1992). This plasticity in ovariole number in Psithyrus has been hypothesized to enable parasitic females to quickly lay a high number of eggs after entering host nests (Richards 1994).
Data is lacking on the nature of chemical signals of reproductive dominance in parasites of complex social insects. Some inquiline ants, such as Leptothorax kutteri or Teleutomyrmex schneideri, have been observed specifically grooming host queens at a very high frequency (Franks et al. 1990; Vander Meer and Alonso 1998) most probably to acquire host queen dominance signals. While parasites of complex social insects are not expected to physically interact with every worker, they might still be able to minimize worker investment in reproduction using physical punishments of dominant workers or indirectly through oophagy (Wenseleers and Ratnieks 2006). Host queens mark their eggs to ensure workers selectively destroy only worker-laid eggs, a strategy we might expect social parasite females to mimic (Foster and Ratnieks 2001). As evidence for this, the hydrocarbon profile of the eggs of the socially parasitic ant Polyergus breviceps closely matches the hydrocarbon profile of host eggs (Johnson et al. 2005).
Host Queen Tolerance
The type of interaction between social parasites and their host queens may have an important impact on parasite dominance strategy and could explain the apparent inability to suppress host worker ovarian development in some social parasites. Social parasite species are usually queen-intolerant and kill the host queen after successful usurpation (Alford 1975; Brandt et al. 2005; Cervo 2006), however some species of social parasite cohabit with the host queen and allow host reproduction to a certain extent (Cervo and Lorenzi 1996; Greene et al. 1978; Fisher 1983, 1988; Lhomme et al. 2013).
Queen-intolerant parasite species, like the socially parasitic wasp Vespula austriaca and parasitic bumble bee B. citrinus, typically fully inhibit the worker ovarian development of their hosts (Fisher 1984; Reed and Akre 1983). By killing or ousting the host queen, queen-intolerant social parasites also remove the host queen’s inhibitory effect on host worker reproduction and thus these parasites must be able to actively match host queen fertility signaling or otherwise control workers. On the other hand, queen-tolerant parasitic females are often unable to fully inhibit host worker ovarian development and may need the presence of the host foundress to maintain reproductive control over workers. This has been observed in the queen-tolerant socially parasitic wasp Polistes atrimandibularis (Cervo and Lorenzi 1996), and also in the wasp Dolichovespula antica, which are unable to inhibit worker ovarian development upon the death of the host queen (Greene et al. 1978; Jeanne 1977). Similarly, in bumble bees the queen-tolerant parasite Bombus ashtoni is incapable of controlling worker ovarian development in the absence of host queens and controls host reproduction by selectively killing host eggs and larvae (Fisher 1984). The author hypothesized B. ashtoni might be able to mask its reproductive status, perhaps by ceasing production of fertility-linked compounds, to avoid queen or worker policing (Fisher 1984). The queen-tolerant ant social parasite Strongylognathus testaceus exhibits another strategy, where the parasitic female allows queen presence but, while allowing worker production, pheromonally inhibits production of the host sexual brood (Guillem et al. 2014).
Although queen-tolerant and queen-intolerant species both use aggressive behavior to maintain their dominant status, aggression has been found to be more intense and frequent in queen-intolerant species in both bumble bees and paper wasps (Cervo 2006; Küpper and Schwammberger 1995; Lhomme et al. 2013; van Honk et al. 1981). For example, queen-intolerant B. vestalis females exhibit a more dominant interaction behavior than queen-tolerant B. sylvestris females (Küpper and Schwammberger 1995; Lhomme et al. 2013), while both being specialized on their respective host. Increased aggression of queen-intolerant species could be an enhanced way to ensure worker control in the absence of queen signals. Inversely, another explanation for this could be a shift in host aggression with parasitic strategy. As the only way for queenless colonies to gain fitness is through host worker reproduction, hosts parasitized by queen-intolerant parasite females should be under strong selection to become more aggressive towards parasites and/or to refine their abilities to discriminate the invader. More aggression is likely to be effective as behavioural experiments in ants have demonstrated that increased aggression towards social parasites results in more host reproduction (Foitzik et al. 2001). In colonies where both host queen and parasitic female reproduce, reciprocal selective pressures should be weaker because host reproductive success is not eliminated and thus host aggression reduced. As such, parasites may evolve increased queen tolerance if aggression limits the potential for successful invasion (Brandt et al. 2005).
The ability to parasitize multiple hosts or to specialize on one (generalist vs. specialist) might also impact, or inversely be impacted by, the social parasite dominance strategy. Although limited data are available on this, especially in complex eusocial insects, we can build some hypotheses from what we know in paper wasps and bumble bees. In paper wasps, P. sulcifer is a specialist parasitizing P. dominula only, P. semenowi is considered semi-generalist parasitizing both P. dominula and P. nimphus, and P. atrimandibularis is a broad generalist on P. dominula, P. nimphus and the more distantly related P. gallicus, P. biglumis and P. associus (Fanelli et al. 2001). These 3 social parasite species differ in the type of interaction they share with their host: the specialist and queen intolerant P. sulcifer and semi-generalist queen tolerant P. semenowi exhibit aggressive interactions towards their hosts (Cervo et al. 1990a; Turillazzi et al. 1990; Zacchi et al. 1996) whereas the queen tolerant broad generalist P. atrimandibularis is a pacifist, avoiding or being submissive to frequent attacks from the hosts (Cervo et al. 1990b). In bumble bees, certain species parasitize only one host (e.g., B. vestalis; van Honk et al. 1981), some are semi-generalists on a few close related host species (e.g., B. bohemicus and B. barbutellus; Kreuter et al. 2012; Lecocq et al. 2011), and some are broad generalists parasitizing distantly related hosts (e.g., B. insularis; Williams et al. 2014). More behavioral data are needed to see any clear trends, but specialist parasitic bumble bees like B. rupestris, B. vestalis and B. citrinus are known to exhibit aggressive behavior towards their host (Fisher 1984; Sladen 1912; Plath 1934; van Honk et al. 1981), whereas the semi-generalist queen-tolerant B. ashtoni exhibits low aggression against its hosts and is suggested to rely on the host queen to control workers (Fisher 1983). Furthermore, the semi-generalist queen-intolerant B. bohemicus has been suggested not to use physical intimidation but to use physical interactions as a mean to transfer inhibitory chemicals to workers (Kreuter et al. 2012).
Predicting Social Parasite Reproductive Dominance Strategies
In simple eusocial insect hosts (Fig. 2a), generalist/queen-tolerant social parasite species, like P. atrimandibularis, are expected to mainly rely on the host queen to control host workers and may have evolved ways to reduce or cease fertility signaling to avoid policing by the workers. Generalist/queen-intolerant social parasite species, like B. insularis, are hypothesized to rely on behavior to take over reproductive control in the nest using high frequency of aggressive behaviors to dominate the host workers. Specialist/queen-tolerant social parasite species, like B. sylvestris, are not expected to be under strong selection to mimic host queen fertility signals but might have evolved an imperfect or “good enough” signal to be accepted as equal of the host queen. Finally, specialist/queen-intolerant social parasite species, such as B. vestalis and P. sulcifer are expected to use both agressive behavior and mimicry of host queen fertility signal, as these species should be under more intense coevolutionary dynamics with their hosts.
In complex eusocial insect hosts (Fig. 2b), social parasites are not expected to make use of aggressive behaviors and thus may not exhibit morphological adaptations to control worker reproduction. Social parasites must instead rely heavily on chemical control of reproduction. Generalist/queen-tolerant social parasite species of complex eusocial hosts, such as Formicoxenus nitidulus, are expected to use a stealthy strategy and avoid using aggressive interactions, as they can rely on the host queen pheromones to control workers. They might even be selected to hide their reproductive status by lowering their fertility signaling to avoid queen and worker policing. Many of these species are able to care for their own brood and seem to exploit host workers mainly to obtain regurgigated food through trophalaxis (Buschinger 2009). Generalist/queen-intolerant species are not found in nature to our knowledge as queen-intolerant inquilines have been observed to specialize on one or two hosts. If they exist, we would predict these species to evolve chemical mimicry of host queen pheromone if a conserved queen pheromone is used across multiple related hosts. Specialist/queen-tolerant social parasites, like Leptothorax kutteri, should not be under strong selection to evolve fine mimicry and may instead rely on chemical acquisition of host queen dominance signals through grooming (Franks et al. 1990) or rely directly on the host queen to control workers of their specific host. Finally, specialist/queen-intolerant social parasites, such as Leptothorax goesswaldi (Buschinger and Klump 1988), are expected to be under strong selection to evolve fine tuned mimicry of the host queen pheromone. Testing these predictions, however, awaits a better understanding of reproductive dominance signals across social lineages and their parasites.
Host-Parasite Coevolution in Fertility Signaling and Other Dominance Signals
There is obviously an important asymmetry regarding the strength of the reciprocal selection pressures between social parasites and their hosts. The reproductive success of a social parasite depends entirely on its interaction with a host colony, and thus selection pressures for successful usurpation and worker control are strong whereas hosts have a lot of other competing, and perhaps countering, selection pressures. There is evidence, however, of an on-going co-evolutionary arms races between social parasites and their hosts. In response to parasitism, hosts have evolved several counter-defensive traits to stave off social parasites in early stages of nest invasion (Fig. 1) including chemical adaptations like diversification of CHC recognition signals in parasitized populations of Formica fusca (i.e., increased number of C25-dimethylalkanes isomers; Martin et al. 2011), behavioral adaptations such as increased aggressiveness in parasitized populations of Temnothorax longispinosus (Pamminger et al. 2011) or even morphological adaptations such as increased body size in parasitized populations of Polistes dominula (Ortolani and Cervo 2010). They may also have evolved counter-measures later in the invasion process to prevent parasite reproductive dominance. For example, host workers of the wasp P. dominula and the ants Formica fusca and F. lemani have enhanced ovary development and/or lay eggs more rapidly in parasitized nests (Chernenko et al. 2011; Cini et al. 2014). Workers may also be able to detect and selectively destroy eggs of their social parasites, as shown in the ants F. fusca and F. lemani parasitized by Formica truncorum (Chernenko et al. 2011) and several species of Temnothorax ants enslaved by Protomognathus americanus (Achenbach and Foitzik 2009).
Eusocial insects produce honest chemical signals of fertility that keep workers informed of queen reproductive status (Keller and Nonacs 1993; Heinze and d’Ettorre 2009), thus allowing optimization of reproductive strategies (Keller and Nonacs 1993; Monnin 2006; Dapporto et al. 2007). The social parasite fertility signal, however, is dishonest by definition. Signal exploitation by the social parasite should lead to additional counter-measures by the hosts involving shifts in the production and response to fertility signals. Data on parasite fertility signaling are limited but some evidence, although still debated (Amsalem et al. 2015; Holman et al. 2017; Smith et al. 2015a), suggests a common class of queen pheromones are used across complex eusocial insect lineages (i.e., long-chain linear and methyl-branched alkanes; Holman et al. 2010, 2013, 2016; Van Oystaeyen et al. 2014). A common pheromone is maladaptive with regard to deterring social parasites. In such case its possible that resistance to parasitism takes place in other steps of nest invasion, such as through strengthened front-line defenses (step 1 and 2, Fig. 1). This could be advantageous as the adaptive benefits of diversification in reproductive control signals may be outweighed by the complex physiological changes they would require. However, some evidence suggests that fertility signals may be more diverse in complex social insects and thus be harder to manipulate. In Temnothorax ants, for example, queen CHC profiles differ more strongly among species than worker profiles and queens are not able to fully repress ovarian development of other species (Brunner et al. 2011).
Physiological constraints related to the mode of response of workers, especially in complex eusocial insects, could explain why queen fertility signals are so conserved over evolutionary time (Holman et al. 2010; Smith et al. 2009, 2012). Several studies in ants suggest that worker response to queen pheromone is innate, with queen pheromone having a primer effect on worker sterility (Holman et al. 2010; Holman et al. 2013). An innate response to a specific queen pheromone might require a specific tunning of the odorant receptors involved, thus evolutionary changes in queen pheromones might be rare because of the necessity for compensatory changes in worker olfactory receptors (Holman et al. 2013; Niehuis et al. 2013). In contrast, in more simple eusocial insects, fertility signaling seems to be interpreted within a learned chemical context. In Odontomachus brunneus the queen pheromone (Z)-9-nonacosene did not inhibit worker reproduction outside the nest (Smith et al. 2015a). In such species we expect that chemical diversification of fertility signaling would be less constrained (Smith and Liebig 2017).
While there is evidence that host CHCs diversify upon parasitism (Martin et al. 2011), there is currently no evidence for evolutionary shifts of fertility signal composition in response to parasitism. Much more work is needed to identify the compounds involved in fertility signaling in hosts and social parasites before this hypothesis can be tested. It would be interesting to compare populations or species of hosts with or without a history of exposure to determine if there is more diversity and variance in fertility compounds in groups with exposure to social parasites.
Social parasites might also exploit a “loophole of honesty” in host fertility signaling. If the “honesty” of host queen fertility signaling is a result of by-products of ovarian development (Holman 2012; Peeters and Liebig 2009) then the signal might represent an indicator that neither queens nor workers can cheat (Holman 2012; Peso et al. 2015). As social parasite females develop their ovaries, they can gain instant dishonest acceptance through this route. Mimicry of host queen pheromones might even be enough to be accepted as a colony member in certain host species regardless of their other cuticular hydrocarbons. In some complex eusocial insects with high worker-queen reproductive dimorphism, queen pheromone can override perception of non-nestmate status and be interpreted in isolation from the learned chemical context (reviewed in Smith and Liebig 2017). In fully grown colonies of the ant Camponotus floridanus, workers are aggressive towards young foreign queens who exhibit a reduced amount of fertility signaling but they accept readily mature foreign queens (Moore and Liebig 2010a, b) suggesting a neurosensory adapted response to the specific queen signals that can outweigh nestmate recognition cues (Smith and Liebig 2017). Host workers might thus be caught in an evolutionary sensory trap with no other ways but to respond positively to mimicked parasite dominance cues and to work against their own reproductive interests.
In simple eusocial insects, we might expect social parasites to have reduced capacity to control workers using a single high abundance chemical signal and hosts can thus avoid parasitism through diversification of fertility signals. In these insects (e.g., Bombus impatiens; Amsalem et al. 2015; Padilla et al. 2016), however, aggressive behaviors likely play a major role in worker inhibition. In these cases, it might be less costly for the hosts to evolve increased physical resistance to parasites than to evolve chemical changes. Some populations of the paper wasp Polistes dominula for example have evolved foundresses that are so large and aggressive that they can successfully defend their nest against social parasites (Ortolani and Cervo 2010).
Other Social Parasite Strategies to Maintain Reproductive Supremacy
Morphological adaptations of social parasites such as thickened cuticle and mandibles (Cervo 1994; Fisher and Sampson 1992; Ondricek-Fallscheer 1992) seem to be mainly used to resist host queen/foundress and worker attacks during invasions. This is especially the case in socially parasitic bumble bees that usually face attacks of multiple workers upon intrusion. For Polistes, which enter host nests and replace the dominant foundress before worker emergence (Cervo 2006), morphological adaptations could help maintain reduced aggressive interactions and dominance throughout the colony cycle. Another strategy to reduce aggression in a nest is to eliminate those individuals most predisposed to it. Workers with developed ovaries are usually the most aggressive and might later on compete for reproduction. To avoid aggression, the bumble bee social parasite Bombus vestalis is able to discriminate and selectively kill host workers with developed ovaries based on the qualitative and quantitative variation of their CHC profile (mainly alkenes and alkanes; Sramkova and Ayasse 2009).
Some social parasites are also able to regulate host activity to their own benefit by increasing brood care behaviors by the hosts (Fucini and Lorenzi 2004). For example, colonies of Polistes biglumis parasitized by P. atrimandibularis exhibit increased foraging and reduced rest (Fucini et al. 2014). As the mechanism of regulation involved is not clearly understood, it would be interesting to test if chemical cues produced by the parasite female could be regulating host activity. Social parasite larvae may also actively divert host feeding behavior to their own benefit. Larvae of the social parasite P. sulcifer seem able to manipulate host workers to increase feeding of the parasite brood (Cervo et al. 2004). In ants, larvae of the social parasite Atemeles pubicollis also manage to obtain greater amounts of food from workers than larvae of their host Formica polyctena (Hölldobler and Wilson 1990). The authors suggested that chemical signals emitted by the brood itself were involved but further investigations are needed to confirm this.
Regulation of host brood populations by social parasites might also be of importance in controlling worker reproduction. Social parasites are often observed destroying eggs and very young larvae of their host, while retaining older larvae and pupae (Cervo 1990; Fisher 1984; Lhomme et al. 2013; Turillazzi and Cervo 1996). This behavior can be seen as a strategy to maximize host care directed towards the parasitic brood while at the same time minimizing destruction of the future workforce. Socially parasitic females may allow host reproduction as an incentive to continue supporting parasite offspring (Cini et al. 2014). Host brood could also be necessary for the social parasite female to fully control host workers. For example, in Polistes dominula, queen dominance signalling is not enough to control worker reproduction which relies in part on the presence of eggs and larvae in the nest cells (Liebig et al. 2005; Monnin et al. 2009).
Finding commonalities in how reproductive dominance occurs across the independent social origins in insects is not necessarily straightforward. Several factors can impact the modalities of reproductive dominance signaling, with the level of social organization being a central one. Through study of how socially parasitic invaders ensure reproductive control and social dominance of another species we can understand the effectiveness of aggression versus chemical control strategies in social systems. We can compare the mechanisms of reproductive control exhibited by host queens in their nests to mechanisms utilized by parasites on those same workers to determine the most essential elements of reproductive control, with social parasites serving essentially as an independent replicate. In cases where chemical control is evident, we have the potential to examine which compounds are more similar between parasite and host to understand key components of the fertility signals for reproductive dominance, results that can be compared more broadly to examine the evolution of chemical signaling for social dominance. As this review has emphasized, we can also look beyond what is shared to what might be different, to appreciate the various routes to social control and parasitism and the factors that might impact them.
This review merges disparate knowledge across diverse social parasitic lineages and has identified some common trends in the modes by which social parasites gain social control. This provides a hypothetical framework to make predictions on how types of sociality and parasitism might affect the mode of reproductive dominance, which can be tested in future research. To this discussion we also bring a coevolutionary perspective, highlighting the different reciprocal evolutionary dynamics between host and parasite, and through which extant parasites have managed to successfully cheat the system. Considering the push-and-pull of selection between hosts and parasites is important to consider when examining the strategies used for each social parasite.
Although trends have emerged, social parasite strategies remain poorly understood. Vital to testing these hypothesis is a need for improved understanding of the chemical components and diversification of host fertility signaling. More comparative approaches to uncover these chemical cues and how they are generated is central to testing the proposed framework. We hope this review encourages these further avenues and particularly pursuit of research on social parasites to address these questions.
We would like to acknowledge Etya Amsalem and Christina Grozinger for helpful discussion on this topic and to Li Tian, Sarthok Rahman, Briana Ezray, Shelby Kilpatrick, and Guillaume Ghisbain, and the external reviewers for many helpful comments. This research was enabled by funding through the Penn State Eberly College of Sciences.
- Alford DV (1975) Bumblebees. Davis-Poynter Ltd, LondonGoogle Scholar
- Amsalem E, Orlova M, Grozinger CM (2015) A conserved class of queen pheromones? Re-evaluating the evidence in bumblebees (Bombus impatiens). Proc R Soc Lond B Biol Sci 282:20151800Google Scholar
- Buschinger A (2009) Social parasitism among ants: a review (Hymenoptera: Formicidae). Myrmecol News 12:219–235Google Scholar
- Camazine S (2003) Self-organization in biological systems. Princeton University Press, PrincetonGoogle Scholar
- Cervo R (1990) II parassitismo sociale nei Polistes (Hymenoptera,Vespidae). Doctorate Thesis, University of FlorenceGoogle Scholar
- Cervo R (2006) Polistes wasps and their social parasites: an overview. Ann Zool Fenn 43:531–549Google Scholar
- Cervo R, Lorenzi MC, Turillazzi S (1990a) Non aggressive usurpation of the nest of Polistes biglumis bimaculatus by the social parasite Sulcopolistes atrimandibularis (Hymenoptera, Vespidae). Insect Soc 37:333–347Google Scholar
- Cervo R, Lorenzi MC, Turillazzi S (1990b) Sulcopolistes atrimandibularis, social parasite and predator of an Alpine Polistes (Hymenoptera, Vespidae). Ethology 86:71–78Google Scholar
- Cervo R, Macinai V, Dechigi F, Turillazzi S (2004) Fast growth of immature brood in a social parasite wasp: a convergent evolution between avian and insect cuckoos. Am Nat 164:814–820Google Scholar
- Emery C (1909) Über den Ursprung der dulotischen, parasitischen und myrmekophilen Ameisen. Biol Cent 29:352–362Google Scholar
- Fisher RM (1983) Inability of the social parasite Psithyrus ashtoni to suppress ovarian development in workers of Bombus affinis (Hymenoptera : Apidae). J Kansas Entomol Soc 56:69–73Google Scholar
- Foitzik S, Deheer CJ, Hunjan DN, Herbers JM (2001) Coevolution in host-parasite systems: behavioural strategies of slave-making ants and their coevolution in host-parasite systems: behavioural strategies of slave-making ants and their hosts. Proc R Soc Lond B Biol Sci 268:1139–1146CrossRefGoogle Scholar
- Fucini S, Lorenzi M (2004) Behavioural counter-adaptations to social parasites in Polistes biglumis, host of P. atrimandibularis Hymenoptera, Vespidae. Insect Soc Life 5:27–29Google Scholar
- Greene A, Akre RD, Landolt PJ (1978) Behavior of the yellowjacket social parasite, Dolichovespula arctica (Rohwer) (Hymenoptera: Vespidae). Melanderia 29:1–28Google Scholar
- Heinze J, d’Ettorre P (2009) Honest and dishonest communication in social Hymenoptera. J Exp Biol 212:1775–1779Google Scholar
- Jeanne RL (1977) Behavior of the obligate social parasite Vespula arctica (Hymenoptera: Vespidae). J Kansas Entomol Soc 50:541–557Google Scholar
- Monnin T (2006) Chemical recognition of reproductive status in social insects. Ann Zool Fennici 43:515–530Google Scholar
- Ondricek-Fallscheer RL (1992) A morphological comparison of the sting apparatuses of socially parasitic and nonparasitic species of yellowjackets (Hymenoptera: Vespidae). Sociobiology 20:245–293Google Scholar
- Peeters C, Liebig J (2009) Fertility signaling as a general mechanism of regulating reproductive division of labor in ants. In: Gadau J, Fewell J (eds) In Organization of Insect Societies: from genome to Sociocomplexity. Harvard University Press, Harvard, pp 220–242Google Scholar
- Plath OE (1934) Bumblebees and their ways. The Macmillan, New YorkGoogle Scholar
- Reed HC, Akre RD (1983) Comparative Colony behavior of the Forest Yellowjacket, Vespula acadica (Sladen) (Hymenoptera: Vespidae). J Kansas Entomol Soc 56:581–606Google Scholar
- Richards KW (1994) Ovarian development, ovariole number, and relationship to body size in Psithyrus spp. (Hymenoptera: Apidae) in southern Alberta. J Kansas Entomol Soc 67:156–168Google Scholar
- Schmid-Hempel P (1998) Parasites in social insects. Princeton University Press, PrincetonGoogle Scholar
- Turillazzi S, Cervo R (1996) Oofagy and infanticide in colonies of social wasps. In: Infanticide and parental care. Harwood Academic Publishers, Amsterdam, pp 213–236Google Scholar
- Vander Meer RK, Alonso LE (1998) Pheromone directed behavior in ants. In: Pheromone communication in social insects: ants, wasps, bees and termites. Westview press, Boulder, pp 159–192Google Scholar
- Williams PH, Thorp RW, Richardson L, Colla S (2014) Bumble bees of North America: an identification guide. Princeton University Press, PrincetonGoogle Scholar
- Wilson EO (1971) The insect societies. Harvard University Press, CambridgeGoogle Scholar
- Zacchi F, Cervo R, Turillazzi S (1996) Polistes semenowi, obligate social parasite, invades the nest of its host, Polistes dominulus (Hymenoptera, Vespidae). Ins Soc Life 1:125–130Google Scholar