The role of juvenile hormone in regulating reproductive physiology and dominance in Dinoponera quadriceps ants

Introduction

In many group-living animals, reproduction is not equally distributed among the breeding members. This phenomenon, known as reproductive skew, occurs in a wide range of communally-breeding species including birds, fishes and insects (Jamieson, 1997; Cuvillier-Hot et al., 2004; Neff, Pitcher & Ramnarine, 2008). Due to its fundamental implications for both ecological and evolutionary processes, this topic has attracted attention over the past decades from both a theoretical (see Kokko & Johnstone, 1999; Johnstone, 2000; Kokko, 2003) and empirical perspective in a variety of study systems (Bourke, Green & Bruford, 1997; Field et al., 1998; Reeve & Keller, 2001; Widdig et al., 2004).

Social insects, in particular ants, have emerged as important model systems to test some of the main predictions of reproductive skew theory due to their wide range of social complexity and life history strategies (Reeve & Keller, 2001). The majority of ant species are highly eusocial, with complex societies in which one or several queens produce all female offspring and her unmated daughter workers perform all other tasks such as brood care, foraging and nest defence (Hölldobler & Wilson, 1990). In such systems, as a result of the haplodiploid sex determination in Hymenoptera, workers are only able to produce male offspring and conflicts between queens and workers, or between workers, arise over male parentage (Ratnieks & Reeve, 1992; Ratnieks, Foster & Wenseleers, 2006). This group conflict and how it is resolved has generated a plethora of ground-breaking work, revolutionizing our understanding of group formation, conflict and maintenance (Ratnieks, Foster & Wenseleers, 2006).

While most modern ants have a specialized queen caste, some genera, such as Dinoponera, have secondarily reverted to simple, queenless societies in which reproduction is monopolized instead by mated, reproductively active workers called gamergates (Peeters, 1997). In most gamergate systems, all workers have the potential to become the dominant reproductive, resulting in strong within-group conflict over reproduction (Peeters, 1997). Such conflicts are often resolved via aggressive behavioural interactions that establish a dominance hierarchy in which only a single, or a small group of workers go on to reproduce. The question of how such ritualized physical aggression is physiologically translated into stable reproductive hierarchies with lower ranked workers not only remaining reproductively inactive but also assuming helper roles, remains poorly understood. In some species such as D. quadriceps, subordinate workers play a role in stabilizing the dominance hierarchy (Monnin & Peeters, 1999; Monnin et al., 2002).

Hormones are prime candidates for the proximate mechanisms underlying this process, because they not only physiologically integrate social stimuli including stress, but also regulate numerous other essential processes in adult insects such as reproduction, maternal behaviour and aggression (Nijhout, 1998; Sasaki, Yamasaki & Nagao, 2007; Tibbetts & Huang, 2010). There are several hormones that have been implicated in this, including dopamine, ecdysone and vitellogenin, with juvenile hormone (JH) being perhaps the best studied. However, it appears that JH can have contrasting effects in different taxa. In primitively eusocial species, such as paper wasps and bumblebees, JH is gonadotropic, stimulating ovary development in the same way as in solitary insects and resulting in individuals being more socially dominant, while in the highly eusocial honeybee, in contrast, JH has lost its gonadotropic effect and instead is involved in regulating division of labour (Robinson & Vargo, 1997). In highly eusocial Lasius niger ants, the gonadotropic effect of JH has not only been lost but reversed, with higher JH levels being associated with reduced egg production (Pamminger et al., 2016; Pamminger, Treanor & Hughes, 2016). In addition JH has been shown to trigger foraging behaviour in some ants, making it a possible candidate to coordinate not only reproductive division of labour, but also division of labour between workers (Robinson & Vargo, 1997; Norman & Hughes, 2016). This might suggest a relatively simple switch in the action of JH with the evolution of complex eusocial societies, but it appears that the evolution dynamics of JH mode of action is more complex than that. In Solenopsis and Pogonomyrmex, ant genera with complex societies, JH exhibits stimulatory functions during reproduction (Brent & Vargo, 2003; Libbrecht et al., 2013). In Streblognathus and Diacamma, ants with simple, queenless societies, low JH titres in gamergates correlates with high individual ranks within the hierarchy and JH application will result in a loss of the reproductive status of the alpha (Sommer, Hölldobler & Rembold, 1993; Cuvillier-Hot, Lenoir & Peeters, 2004; Brent et al., 2006). In Harpegnathos ants, which also have simple, societies, where gamergates can reproduce following the founding queen’s death, JH levels do not differ between reproductive and non-reproductive individuals, and experimental elevation of JH levels through the application of Juvenile Hormone analogue had no effect on egg production (Penick, Liebig & Brent, 2011).

Here we investigate the effects of JH on reproductive physiology and social dominance in the queenless ponerine ant Dinoponera quadriceps, by using topical application of the JH analogue (JHa) methoprene to experimentally manipulate JH levels. This species is of particular interest because it is one of ca. 100 species to have undergone an evolutionary reversion from a highly eusocial ancestor with a queen caste back to its basal state with queenless, simple societies (Peeters, 1997; Monnin & Peeters, 1998). All females in D. quadriceps are morphologically identical, with a single dominant gamergate, the alpha, actively suppressing a group of the higher ranked workers from becoming reproductively active with ritualized physical aggression including antennal blocking and boxing (Monnin & Peeters, 1998; Grainger et al., 2014). The presence of an alpha within the colony not only inhibits ovary activation in workers, the first step towards becoming reproductively active, but also results in submissive behaviour by subordinates (Smith et al., 2011; Asher et al., 2013). The physiological phenotypic differences between alphas and subordinates result from subtle differences in transcriptional network organisation, involving both conserved and novel genes (Patalano et al., 2015). If JH functions similarly to other queenless ant species, then we predict JH will suppress ovarian development and cause high ranking workers to decrease in status. If JH functions as a gonadotropin, similar to solitary insects and social wasps, then we predict JH to activate ovarian development and potentially to move up in the hierarchy. We focus for our experimental manipulation on high-ranked, but not alpha, workers because these have both the potential to move up the hierarchy to become reproductives and the potential to lose their position in the social hierarchy and become middle or low-ranked workers. We measure the effect of JH manipulation on ovary development and dominance behaviour. If JH links reproduction and hierarchy-related behaviours it would then provide a proximate physiological explanation for rank-associated trait syndromes.

Methods

We used 13 colonies of D. quadriceps, which were collected from Bahia state, Brazil in November 2014 under permit from Instituto Brasileiro do Meio Ambiente e dos Recursos Naturais (IBAMA; 14BR004553). All colonies were maintained in the lab at 27 °C and 80% relative humidity for at least six months before the experiment. Colonies were fed with Tenebrio molitor larvae and apple, and had ad libitum access to water. Each individual was uniquely marked on the pronotum with numbered tags.

Results

The experimental ants did not differ in size or weight between treatments or colonies (respectively: Pseudo F_1,31 = 0.35, P = 0.72; Pseudo F_11,31 = 1.04, P = 0.4; Table 1), and worker fertility also did not differ between colonies (Pseudo F_11,31 = 1.82, P = 0.11). However, worker fertility was affected by treatment, with JHa-treated individuals being less fertile compared to those treated with acetone control under all measures of fertility taken here (oocyte number, average ovariole width, minimum ovariole width and maximum ovariole width; 22; Tables 1 and 2; Fig. 1). The CoA ants were more variable in fertility than the JHa group (Fig. 1). Furthermore, there were smaller differences in fertility within the JHa group (Fig. 1). In contrast to the effects of JHa treatment on worker fertility, we found no significant differences between treatment or colonies on social dominance behaviour (respectively: Pseudo F_1,31 = 0.74, P = 0.6; Pseudo F_11,31 = 1.37, P = 0.06; Fig. 2).

Discussion

Our results show that JH has a role in regulating reproduction in D. quadriceps. Experimentally elevated JH levels not only decreased the number of vitellogenic eggs in high-ranked workers, but also resulted in an overall decrease in the size of individual ovarioles, indicating a substantial reduction in reproductive potential. In solitary insects JH often has the opposite effect by stimulating the production of vitellogenic oocytes and the same is true for many primitively eusocial insects, such as non-swarm founding wasps and bumblebees (Robinson & Vargo, 1997). It has been suggested that the functional reversal of JH in reproduction was important in the evolution of complex societies, and a large number of studies demonstrate often radical changes in the regulatory architecture of reproduction in eusocial species (Robinson & Vargo, 1997; Hartfelder, 2000; Bloch et al., 2009). The classic example for this argument is the remodelling of the regulatory function of JH in honeybees and some ants (Robinson & Vargo, 1997; Bloch et al., 2009; Pamminger et al., 2016; Azevedo et al., 2016; Pamminger, Treanor & Hughes, 2016), however a small number of studies clearly indicates that high social organization is possible without it (e.g., Brent & Vargo, 2003; Kelstrup, Hartfelder & Wossler, 2015). D. quadriceps supports the former findings by demonstrating that a remodelling of JH function which inhibits reproduction can also be associated with simple social organization. This makes sense given the evolutionary position of D. quadriceps, with Dinoponera having secondarily reverted to simple, queenless societies from a highly eusocial ancestor (Peeters, 1997; Monnin & Peeters, 1998; Monnin & Peeters, 1999). This further supports the notion that there is likely no causal link between the remodelling of the JH function in reproduction and the organisational complexity of insect societies.

The results presented here, that JH significantly decreases fertility, are therefore consistent with those for other gamergate-led ant societies (Harpegnathos, Streblognathus and Diacamma; Sommer, Hölldobler & Rembold, 1993; Cuvillier-Hot, Lenoir & Peeters, 2004; Penick, Liebig & Brent, 2011). The role of JH in such societies is thus relatively clear, but there remain questions on how JH functions in true queens of other species. Importantly in these societies JH appears to be involved as part of an ‘honest’ signal that informs other colony members about their fertility (Cuvillier-Hot, Lenoir & Peeters, 2004), a key cue for the maintenance of dominance hierarchies in these societies. In studies on a gamergate ant society, a social wasp and a termite, topical applications of JH affect cuticular hydrocarbon (CHC) profiles of adult and larval individuals such that the profile becomes more ‘reproductive-like’ and is perceived so by colony members (Kelstrup et al., 2014; Brent et al., 2016; Penick & Liebig, 2017). However in Streblognathus ants alphas treated with JH had their fertility reduced and the CHC profile matched more closely with that of a sterile worker than an alpha (Cuvillier-Hot et al., 2004)

In contrast to the relatively well-studied effects of JH on reproductive physiology, little is known about the regulatory role of JH in behaviour for most insects. The association between JH and aggression, maternal behaviour and activity has been documented in insects (Nijhout, 1998; Pearce, Huang & Breed, 2001; Tibbetts & Izzo, 2009; Tibbetts, Vernier & Jinn, 2013), but these studies are restricted to only a handful of species. In the honey bee Apis mellifera, JH, in combination with the yolk precursor vitellogenin, regulates one of the major behavioural transitions in the adult honeybee worker from within-nest behaviour to external foraging (Robinson & Vargo, 1997). This transition is associated with a major remodelling of the behaviour repertoire and indicates the far-reaching regulatory potential of JH in behaviour. A similar function of JH has been documented in Pogonomyrmex californicus harvester ants and Acromyrmex echinator leaf-cutting ants (Dolezal et al., 2009; Norman & Hughes, 2016), demonstrating that JH can generate forager-like behavioural phenotypes. In contrast to our expectations, we find no measurable effects of JH on worker behaviour or position in the hierarchy in D. quadriceps. This could indicate that JH, fertility and dominance are decoupled in D. quadriceps. Indeed, a lack of behavioural effect of JHa on alphas in Streblognathus ponerine ants has been reported previously (Cuvillier-Hot et al., 2004). However, it is more likely that our study simply lacked the power to detect an effect. Our study was deliberately focussed on only high-ranked workers because these were the individuals in which both positive and negative effects could potentially be seen, and it may be that inclusion of ants from the full spectrum of the social hierarchy or using in-nest behavioural observations may reveal effects.

Conclusions

Although JH is now known to have an important (though variable) role in the physiology of reproductive dominance in social insects, other hormones are also likely to be as, or more, important to reproductive status and social dominance in D. quadriceps. In particular, ecdysone or vitellogenin generate observed rank-specific phenotypes in other social insects (Hartfelder, 2000), and dopamine may also play a central role in the regulation of dominance and reproduction in species with simple societies (Sasaki, Yamasaki & Nagao, 2007; Okada et al., 2015; Ohkawara & Aonuma, 2016). Further work combining behavioural, genetic and physiological work is needed to illuminate the regulatory underpinning of reproductive hierarchies in simple ant societies. When looking at the broader phylogenetic picture there is accumulating evidence that JH occupies a stunning range of different, often opposite, regulatory functions. The question of how such incredible regulatory flexibility is possible without compromising fitness-relevant functions is intriguing and a promising target for further molecular and comparative investigations.

Supplemental Information